✨ feat(controller): gracefully handle missing Uptime Kuma configuration

Previously, the uptime status endpoint returned HTTP 400 with
“未配置 Uptime Kuma URL/Slug” when either option was not set, resulting in
frontend error states.

Changes:
• Treat absence of `UptimeKumaUrl` or `UptimeKumaSlug` as a valid scenario.
• Immediately respond with HTTP 200, `success: true`, and an empty `data` array.
• Preserve existing behavior when both options are provided.

This prevents unnecessary error notifications on the dashboard when
Uptime Kuma integration is not configured and improves overall UX.

.github/workflows/docker-image-alpha.yml

@@ -1,14 +1,15 @@
-name: Publish Docker image (amd64)
+name: Publish Docker image (alpha)
 
 on:
   push:
-    tags:
-      - '*'
+    branches:
+      - alpha
   workflow_dispatch:
     inputs:
       name:
-        description: 'reason'
+        description: "reason"
         required: false
+
 jobs:
   push_to_registries:
     name: Push Docker image to multiple registries
@@ -18,37 +19,44 @@ jobs:
       contents: read
     steps:
       - name: Check out the repo
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
 
       - name: Save version info
         run: |
-          git describe --tags > VERSION 
+          echo "alpha-$(date +'%Y%m%d')-$(git rev-parse --short HEAD)" > VERSION
 
       - name: Log in to Docker Hub
-        uses: docker/login-action@v2
+        uses: docker/login-action@v3
         with:
           username: ${{ secrets.DOCKERHUB_USERNAME }}
           password: ${{ secrets.DOCKERHUB_TOKEN }}
 
       - name: Log in to the Container registry
-        uses: docker/login-action@v2
+        uses: docker/login-action@v3
         with:
           registry: ghcr.io
           username: ${{ github.actor }}
           password: ${{ secrets.GITHUB_TOKEN }}
 
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
       - name: Extract metadata (tags, labels) for Docker
         id: meta
-        uses: docker/metadata-action@v4
+        uses: docker/metadata-action@v5
         with:
           images: |
             calciumion/new-api
             ghcr.io/${{ github.repository }}
+          tags: |
+            type=raw,value=alpha
+            type=raw,value=alpha-{{date 'YYYYMMDD'}}-{{sha}}
 
       - name: Build and push Docker images
-        uses: docker/build-push-action@v3
+        uses: docker/build-push-action@v5
         with:
           context: .
+          platforms: linux/amd64,linux/arm64
           push: true
           tags: ${{ steps.meta.outputs.tags }}
-          labels: ${{ steps.meta.outputs.labels }}
+          labels: ${{ steps.meta.outputs.labels }}

.github/workflows/docker-image-arm64.yml

@@ -1,15 +1,9 @@
-name: Publish Docker image (arm64)
+name: Publish Docker image (Multi Registries)
 
 on:
   push:
     tags:
       - '*'
-      - '!*-alpha*'
-  workflow_dispatch:
-    inputs:
-      name:
-        description: 'reason'
-        required: false
 jobs:
   push_to_registries:
     name: Push Docker image to multiple registries
@@ -19,26 +13,26 @@ jobs:
       contents: read
     steps:
       - name: Check out the repo
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
 
       - name: Save version info
         run: |
           git describe --tags > VERSION 
 
       - name: Set up QEMU
-        uses: docker/setup-qemu-action@v2
+        uses: docker/setup-qemu-action@v3
 
       - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v2
+        uses: docker/setup-buildx-action@v3
 
       - name: Log in to Docker Hub
-        uses: docker/login-action@v2
+        uses: docker/login-action@v3
         with:
           username: ${{ secrets.DOCKERHUB_USERNAME }}
           password: ${{ secrets.DOCKERHUB_TOKEN }}
 
       - name: Log in to the Container registry
-        uses: docker/login-action@v2
+        uses: docker/login-action@v3
         with:
           registry: ghcr.io
           username: ${{ github.actor }}
@@ -46,14 +40,14 @@ jobs:
 
       - name: Extract metadata (tags, labels) for Docker
         id: meta
-        uses: docker/metadata-action@v4
+        uses: docker/metadata-action@v5
         with:
           images: |
             calciumion/new-api
             ghcr.io/${{ github.repository }}
 
       - name: Build and push Docker images
-        uses: docker/build-push-action@v3
+        uses: docker/build-push-action@v5
         with:
           context: .
           platforms: linux/amd64,linux/arm64

.github/workflows/linux-release.yml

@@ -3,6 +3,11 @@ permissions:
   contents: write
 
 on:
+  workflow_dispatch:
+    inputs:
+      name:
+        description: 'reason'
+        required: false
   push:
     tags:
       - '*'
@@ -15,16 +20,16 @@ jobs:
         uses: actions/checkout@v3
         with:
           fetch-depth: 0
-      - uses: actions/setup-node@v3
+      - uses: oven-sh/setup-bun@v2
         with:
-          node-version: 18
+          bun-version: latest
       - name: Build Frontend
         env:
           CI: ""
         run: |
           cd web
-          npm install
-          REACT_APP_VERSION=$(git describe --tags) npm run build
+          bun install
+          DISABLE_ESLINT_PLUGIN='true' VITE_REACT_APP_VERSION=$(git describe --tags) bun run build
           cd ..
       - name: Set up Go
         uses: actions/setup-go@v3

.github/workflows/macos-release.yml

@@ -3,6 +3,11 @@ permissions:
   contents: write
 
 on:
+  workflow_dispatch:
+    inputs:
+      name:
+        description: 'reason'
+        required: false
   push:
     tags:
       - '*'
@@ -15,16 +20,16 @@ jobs:
         uses: actions/checkout@v3
         with:
           fetch-depth: 0
-      - uses: actions/setup-node@v3
+      - uses: oven-sh/setup-bun@v2
         with:
-          node-version: 18
+          bun-version: latest
       - name: Build Frontend
         env:
           CI: ""
         run: |
           cd web
-          npm install
-          REACT_APP_VERSION=$(git describe --tags) npm run build
+          bun install
+          DISABLE_ESLINT_PLUGIN='true' VITE_REACT_APP_VERSION=$(git describe --tags) bun run build
           cd ..
       - name: Set up Go
         uses: actions/setup-go@v3

.github/workflows/windows-release.yml

@@ -3,6 +3,11 @@ permissions:
   contents: write
 
 on:
+  workflow_dispatch:
+    inputs:
+      name:
+        description: 'reason'
+        required: false
   push:
     tags:
       - '*'
@@ -18,16 +23,16 @@ jobs:
         uses: actions/checkout@v3
         with:
           fetch-depth: 0
-      - uses: actions/setup-node@v3
+      - uses: oven-sh/setup-bun@v2
         with:
-          node-version: 18
+          bun-version: latest
       - name: Build Frontend
         env:
           CI: ""
         run: |
           cd web
-          npm install
-          REACT_APP_VERSION=$(git describe --tags) npm run build
+          bun install
+          DISABLE_ESLINT_PLUGIN='true' VITE_REACT_APP_VERSION=$(git describe --tags) bun run build
           cd ..
       - name: Set up Go
         uses: actions/setup-go@v3

.gitignore

@@ -9,4 +9,5 @@ logs
 web/dist
 .env
 one-api
-.DS_Store
+.DS_Store
+tiktoken_cache

README.en.md

@@ -1,10 +1,13 @@
+<p align="right">
+   <a href="./README.md">中文</a> | <strong>English</strong>
+</p>
 <div align="center">
 
 ![new-api](/web/public/logo.png)
 
 # New API
 
-🍥 Next Generation LLM Gateway and AI Asset Management System
+🍥 Next-Generation Large Model Gateway and AI Asset Management System
 
 <a href="https://trendshift.io/repositories/8227" target="_blank"><img src="https://trendshift.io/api/badge/repositories/8227" alt="Calcium-Ion%2Fnew-api | Trendshift" style="width: 250px; height: 55px;" width="250" height="55"/></a>
 
@@ -33,171 +36,159 @@
 > This is an open-source project developed based on [One API](https://github.com/songquanpeng/one-api)
 
 > [!IMPORTANT]  
-> - Users must comply with OpenAI's [Terms of Use](https://openai.com/policies/terms-of-use) and relevant laws and regulations. Not to be used for illegal purposes.
-> - This project is for personal learning only. Stability is not guaranteed, and no technical support is provided.
+> - This project is for personal learning purposes only, with no guarantee of stability or technical support.
+> - Users must comply with OpenAI's [Terms of Use](https://openai.com/policies/terms-of-use) and **applicable laws and regulations**, and must not use it for illegal purposes.
+> - According to the [《Interim Measures for the Management of Generative Artificial Intelligence Services》](http://www.cac.gov.cn/2023-07/13/c_1690898327029107.htm), please do not provide any unregistered generative AI services to the public in China.
+
+## 📚 Documentation
+
+For detailed documentation, please visit our official Wiki: [https://docs.newapi.pro/](https://docs.newapi.pro/)
+
+You can also access the AI-generated DeepWiki:
+[![Ask DeepWiki](https://deepwiki.com/badge.svg)](https://deepwiki.com/QuantumNous/new-api)
 
 ## ✨ Key Features
 
-1. 🎨 New UI interface (some interfaces pending update)
-2. 🌍 Multi-language support (work in progress)
-3. 🎨 Added [Midjourney-Proxy(Plus)](https://github.com/novicezk/midjourney-proxy) interface support, [Integration Guide](Midjourney.md)
-4. 💰 Online recharge support, configurable in system settings:
-    - [x] EasyPay
-5. 🔍 Query usage quota by key:
-    - Works with [neko-api-key-tool](https://github.com/Calcium-Ion/neko-api-key-tool)
-6. 📑 Configurable items per page in pagination
-7. 🔄 Compatible with original One API database (one-api.db)
-8. 💵 Support per-request model pricing, configurable in System Settings - Operation Settings
-9. ⚖️ Support channel **weighted random** selection
-10. 📈 Data dashboard (console)
-11. 🔒 Configurable model access per token
-12. 🤖 Telegram authorization login support:
-    1. System Settings - Configure Login Registration - Allow Telegram Login
-    2. Send /setdomain command to [@Botfather](https://t.me/botfather)
-    3. Select your bot, then enter http(s)://your-website/login
-    4. Telegram Bot name is the bot username without @
-13. 🎵 Added [Suno API](https://github.com/Suno-API/Suno-API) interface support, [Integration Guide](Suno.md)
-14. 🔄 Support for Rerank models, compatible with Cohere and Jina, can integrate with Dify, [Integration Guide](Rerank.md)
-15. ⚡ **[OpenAI Realtime API](https://platform.openai.com/docs/guides/realtime/integration)** - Support for OpenAI's Realtime API, including Azure channels
-16. 🧠 Support for setting reasoning effort through model name suffix:
-    - Add suffix `-high` to set high reasoning effort (e.g., `o3-mini-high`)
-    - Add suffix `-medium` to set medium reasoning effort
-    - Add suffix `-low` to set low reasoning effort
-17. 🔄 Thinking to content option `thinking_to_content` in `Channel->Edit->Channel Extra Settings`, default is `false`, when `true`, the `reasoning_content` of the thinking content will be converted to `<think>` tags and concatenated to the content returned.
-18. 🔄 Model rate limit, support setting total request limit and successful request limit in `System Settings->Rate Limit Settings`
-19. 💰 Cache billing support, when enabled can charge a configurable ratio for cache hits:
-    1. Set `Prompt Cache Ratio` in `System Settings -> Operation Settings`
-    2. Set `Prompt Cache Ratio` in channel settings, range 0-1 (e.g., 0.5 means 50% charge on cache hits)
+New API offers a wide range of features, please refer to [Features Introduction](https://docs.newapi.pro/wiki/features-introduction) for details:
+
+1. 🎨 Brand new UI interface
+2. 🌍 Multi-language support
+3. 💰 Online recharge functionality (YiPay)
+4. 🔍 Support for querying usage quotas with keys (works with [neko-api-key-tool](https://github.com/Calcium-Ion/neko-api-key-tool))
+5. 🔄 Compatible with the original One API database
+6. 💵 Support for pay-per-use model pricing
+7. ⚖️ Support for weighted random channel selection
+8. 📈 Data dashboard (console)
+9. 🔒 Token grouping and model restrictions
+10. 🤖 Support for more authorization login methods (LinuxDO, Telegram, OIDC)
+11. 🔄 Support for Rerank models (Cohere and Jina), [API Documentation](https://docs.newapi.pro/api/jinaai-rerank)
+12. ⚡ Support for OpenAI Realtime API (including Azure channels), [API Documentation](https://docs.newapi.pro/api/openai-realtime)
+13. ⚡ Support for Claude Messages format, [API Documentation](https://docs.newapi.pro/api/anthropic-chat)
+14. Support for entering chat interface via /chat2link route
+15. 🧠 Support for setting reasoning effort through model name suffixes:
+    1. OpenAI o-series models
+        - Add `-high` suffix for high reasoning effort (e.g.: `o3-mini-high`)
+        - Add `-medium` suffix for medium reasoning effort (e.g.: `o3-mini-medium`)
+        - Add `-low` suffix for low reasoning effort (e.g.: `o3-mini-low`)
+    2. Claude thinking models
+        - Add `-thinking` suffix to enable thinking mode (e.g.: `claude-3-7-sonnet-20250219-thinking`)
+16. 🔄 Thinking-to-content functionality
+17. 🔄 Model rate limiting for users
+18. 💰 Cache billing support, which allows billing at a set ratio when cache is hit:
+    1. Set the `Prompt Cache Ratio` option in `System Settings-Operation Settings`
+    2. Set `Prompt Cache Ratio` in the channel, range 0-1, e.g., setting to 0.5 means billing at 50% when cache is hit
     3. Supported channels:
         - [x] OpenAI
-        - [x] Azure 
+        - [x] Azure
         - [x] DeepSeek
-        - [ ] Claude
+        - [x] Claude
 
 ## Model Support
-This version additionally supports:
-1. Third-party model **gpts** (gpt-4-gizmo-*)
-2. [Midjourney-Proxy(Plus)](https://github.com/novicezk/midjourney-proxy) interface, [Integration Guide](Midjourney.md)
-3. Custom channels with full API URL support
-4. [Suno API](https://github.com/Suno-API/Suno-API) interface, [Integration Guide](Suno.md)
-5. Rerank models, supporting [Cohere](https://cohere.ai/) and [Jina](https://jina.ai/), [Integration Guide](Rerank.md)
-6. Dify
 
-You can add custom models gpt-4-gizmo-* in channels. These are third-party models and cannot be called with official OpenAI keys.
+This version supports multiple models, please refer to [API Documentation-Relay Interface](https://docs.newapi.pro/api) for details:
 
-## Additional Configurations Beyond One API
-- `GENERATE_DEFAULT_TOKEN`: Generate initial token for new users, default `false`
-- `STREAMING_TIMEOUT`: Set streaming response timeout, default 60 seconds
-- `DIFY_DEBUG`: Output workflow and node info to client for Dify channel, default `true`
-- `FORCE_STREAM_OPTION`: Override client stream_options parameter, default `true`
-- `GET_MEDIA_TOKEN`: Calculate image tokens, default `true`
-- `GET_MEDIA_TOKEN_NOT_STREAM`: Calculate image tokens in non-stream mode, default `true`
-- `UPDATE_TASK`: Update async tasks (Midjourney, Suno), default `true`
-- `GEMINI_MODEL_MAP`: Specify Gemini model versions (v1/v1beta), format: "model:version", comma-separated
-- `COHERE_SAFETY_SETTING`: Cohere model [safety settings](https://docs.cohere.com/docs/safety-modes#overview), options: `NONE`, `CONTEXTUAL`, `STRICT`, default `NONE`
-- `GEMINI_VISION_MAX_IMAGE_NUM`: Gemini model maximum image number, default `16`, set to `-1` to disable
-- `MAX_FILE_DOWNLOAD_MB`: Maximum file download size in MB, default `20`
-- `CRYPTO_SECRET`: Encryption key for encrypting database content
-- `AZURE_DEFAULT_API_VERSION`: Azure channel default API version, if not specified in channel settings, use this version, default `2024-12-01-preview`
-- `NOTIFICATION_LIMIT_DURATION_MINUTE`: Duration of notification limit in minutes, default `10`
-- `NOTIFY_LIMIT_COUNT`: Maximum number of user notifications in the specified duration, default `2`
+1. Third-party models **gpts** (gpt-4-gizmo-*)
+2. Third-party channel [Midjourney-Proxy(Plus)](https://github.com/novicezk/midjourney-proxy) interface, [API Documentation](https://docs.newapi.pro/api/midjourney-proxy-image)
+3. Third-party channel [Suno API](https://github.com/Suno-API/Suno-API) interface, [API Documentation](https://docs.newapi.pro/api/suno-music)
+4. Custom channels, supporting full call address input
+5. Rerank models ([Cohere](https://cohere.ai/) and [Jina](https://jina.ai/)), [API Documentation](https://docs.newapi.pro/api/jinaai-rerank)
+6. Claude Messages format, [API Documentation](https://docs.newapi.pro/api/anthropic-chat)
+7. Dify, currently only supports chatflow
+
+## Environment Variable Configuration
+
+For detailed configuration instructions, please refer to [Installation Guide-Environment Variables Configuration](https://docs.newapi.pro/installation/environment-variables):
+
+- `GENERATE_DEFAULT_TOKEN`: Whether to generate initial tokens for newly registered users, default is `false`
+- `STREAMING_TIMEOUT`: Streaming response timeout, default is 60 seconds
+- `DIFY_DEBUG`: Whether to output workflow and node information for Dify channels, default is `true`
+- `FORCE_STREAM_OPTION`: Whether to override client stream_options parameter, default is `true`
+- `GET_MEDIA_TOKEN`: Whether to count image tokens, default is `true`
+- `GET_MEDIA_TOKEN_NOT_STREAM`: Whether to count image tokens in non-streaming cases, default is `true`
+- `UPDATE_TASK`: Whether to update asynchronous tasks (Midjourney, Suno), default is `true`
+- `COHERE_SAFETY_SETTING`: Cohere model safety settings, options are `NONE`, `CONTEXTUAL`, `STRICT`, default is `NONE`
+- `GEMINI_VISION_MAX_IMAGE_NUM`: Maximum number of images for Gemini models, default is `16`
+- `MAX_FILE_DOWNLOAD_MB`: Maximum file download size in MB, default is `20`
+- `CRYPTO_SECRET`: Encryption key used for encrypting database content
+- `AZURE_DEFAULT_API_VERSION`: Azure channel default API version, default is `2025-04-01-preview`
+- `NOTIFICATION_LIMIT_DURATION_MINUTE`: Notification limit duration, default is `10` minutes
+- `NOTIFY_LIMIT_COUNT`: Maximum number of user notifications within the specified duration, default is `2`
+- `ERROR_LOG_ENABLED=true`: Whether to record and display error logs, default is `false`
 
 ## Deployment
 
+For detailed deployment guides, please refer to [Installation Guide-Deployment Methods](https://docs.newapi.pro/installation):
+
 > [!TIP]
-> Latest Docker image: `calciumion/new-api:latest`  
-> Default account: root, password: 123456
+> Latest Docker image: `calciumion/new-api:latest`
 
-### Multi-Server Deployment
-- Must set `SESSION_SECRET` environment variable, otherwise login state will not be consistent across multiple servers.
-- If using a public Redis, must set `CRYPTO_SECRET` environment variable, otherwise Redis content will not be able to be obtained in multi-server deployment.
+### Multi-machine Deployment Considerations
+- Environment variable `SESSION_SECRET` must be set, otherwise login status will be inconsistent across multiple machines
+- If sharing Redis, `CRYPTO_SECRET` must be set, otherwise Redis content cannot be accessed across multiple machines
 
-### Requirements
-- Local database (default): SQLite (Docker deployment must mount `/data` directory)
-- Remote database: MySQL >= 5.7.8, PgSQL >= 9.6
+### Deployment Requirements
+- Local database (default): SQLite (Docker deployment must mount the `/data` directory)
+- Remote database: MySQL version >= 5.7.8, PgSQL version >= 9.6
 
-### Deployment with BT Panel
-Install BT Panel (**version 9.2.0** or above) from [BT Panel Official Website](https://www.bt.cn/new/download.html), choose the stable version script to download and install.  
-After installation, log in to BT Panel and click Docker in the menu bar. First-time access will prompt to install Docker service. Click Install Now and follow the prompts to complete installation.  
-After installation, find **New-API** in the app store, click install, configure basic options to complete installation.  
-[Pictorial Guide](BT.md)
+### Deployment Methods
 
-### Docker Deployment
+#### Using BaoTa Panel Docker Feature
+Install BaoTa Panel (version **9.2.0** or above), find **New-API** in the application store and install it.
+[Tutorial with images](./docs/BT.md)
 
-### Using Docker Compose (Recommended)
+#### Using Docker Compose (Recommended)
 ```shell
-# Clone project
+# Download the project
 git clone https://github.com/Calcium-Ion/new-api.git
 cd new-api
 # Edit docker-compose.yml as needed
-# nano docker-compose.yml
-# vim docker-compose.yml
 # Start
 docker-compose up -d
 ```
 
-#### Update Version
+#### Using Docker Image Directly
 ```shell
-docker-compose pull
-docker-compose up -d
-```
-
-### Direct Docker Image Usage
-```shell
-# SQLite deployment:
+# Using SQLite
 docker run --name new-api -d --restart always -p 3000:3000 -e TZ=Asia/Shanghai -v /home/ubuntu/data/new-api:/data calciumion/new-api:latest
 
-# MySQL deployment (add -e SQL_DSN="root:123456@tcp(localhost:3306)/oneapi"), modify database connection parameters as needed
-# Example:
+# Using MySQL
 docker run --name new-api -d --restart always -p 3000:3000 -e SQL_DSN="root:123456@tcp(localhost:3306)/oneapi" -e TZ=Asia/Shanghai -v /home/ubuntu/data/new-api:/data calciumion/new-api:latest
 ```
 
-#### Update Version
-```shell
-# Pull the latest image
-docker pull calciumion/new-api:latest
-# Stop and remove the old container
-docker stop new-api
-docker rm new-api
-# Run the new container with the same parameters as before
-docker run --name new-api -d --restart always -p 3000:3000 -e TZ=Asia/Shanghai -v /home/ubuntu/data/new-api:/data calciumion/new-api:latest
-```
+## Channel Retry and Cache
+Channel retry functionality has been implemented, you can set the number of retries in `Settings->Operation Settings->General Settings`. It is **recommended to enable caching**.
 
-Alternatively, you can use Watchtower for automatic updates (not recommended, may cause database incompatibility):
-```shell
-docker run --rm -v /var/run/docker.sock:/var/run/docker.sock containrrr/watchtower -cR
-```
+### Cache Configuration Method
+1. `REDIS_CONN_STRING`: Set Redis as cache
+2. `MEMORY_CACHE_ENABLED`: Enable memory cache (no need to set manually if Redis is set)
 
-## Channel Retry
-Channel retry is implemented, configurable in `Settings->Operation Settings->General Settings`. **Cache recommended**.  
-If retry is enabled, the system will automatically use the next priority channel for the same request after a failed request.
+## API Documentation
 
-### Cache Configuration
-1. `REDIS_CONN_STRING`: Use Redis as cache
-    + Example: `REDIS_CONN_STRING=redis://default:redispw@localhost:49153`
-2. `MEMORY_CACHE_ENABLED`: Enable memory cache, default `false`
-    + Example: `MEMORY_CACHE_ENABLED=true`
+For detailed API documentation, please refer to [API Documentation](https://docs.newapi.pro/api):
 
-### Why Some Errors Don't Retry
-Error codes 400, 504, 524 won't retry
-### To Enable Retry for 400
-In `Channel->Edit`, set `Status Code Override` to:
-```json
-{
-  "400": "500"
-}
-```
-
-## Integration Guides
-- [Midjourney Integration](Midjourney.md)
-- [Suno Integration](Suno.md)
+- [Chat API](https://docs.newapi.pro/api/openai-chat)
+- [Image API](https://docs.newapi.pro/api/openai-image)
+- [Rerank API](https://docs.newapi.pro/api/jinaai-rerank)
+- [Realtime API](https://docs.newapi.pro/api/openai-realtime)
+- [Claude Chat API (messages)](https://docs.newapi.pro/api/anthropic-chat)
 
 ## Related Projects
 - [One API](https://github.com/songquanpeng/one-api): Original project
 - [Midjourney-Proxy](https://github.com/novicezk/midjourney-proxy): Midjourney interface support
-- [chatnio](https://github.com/Deeptrain-Community/chatnio): Next-gen AI B/C solution
-- [neko-api-key-tool](https://github.com/Calcium-Ion/neko-api-key-tool): Query usage quota by key
+- [chatnio](https://github.com/Deeptrain-Community/chatnio): Next-generation AI one-stop B/C-end solution
+- [neko-api-key-tool](https://github.com/Calcium-Ion/neko-api-key-tool): Query usage quota with key
+
+Other projects based on New API:
+- [new-api-horizon](https://github.com/Calcium-Ion/new-api-horizon): High-performance optimized version of New API
+- [VoAPI](https://github.com/VoAPI/VoAPI): Frontend beautified version based on New API
+
+## Help and Support
+
+If you have any questions, please refer to [Help and Support](https://docs.newapi.pro/support):
+- [Community Interaction](https://docs.newapi.pro/support/community-interaction)
+- [Issue Feedback](https://docs.newapi.pro/support/feedback-issues)
+- [FAQ](https://docs.newapi.pro/support/faq)
 
 ## 🌟 Star History
 
-[![Star History Chart](https://api.star-history.com/svg?repos=Calcium-Ion/new-api&type=Date)](https://star-history.com/#Calcium-Ion/new-api&Date)
+[![Star History Chart](https://api.star-history.com/svg?repos=Calcium-Ion/new-api&type=Date)](https://star-history.com/#Calcium-Ion/new-api&Date)

README.md

@@ -7,7 +7,6 @@
 
 # New API
 
-
 🍥新一代大模型网关与AI资产管理系统
 
 <a href="https://trendshift.io/repositories/8227" target="_blank"><img src="https://trendshift.io/api/badge/repositories/8227" alt="Calcium-Ion%2Fnew-api | Trendshift" style="width: 250px; height: 55px;" width="250" height="55"/></a>
@@ -37,199 +36,158 @@
 > 本项目为开源项目，在[One API](https://github.com/songquanpeng/one-api)的基础上进行二次开发
 
 > [!IMPORTANT]  
-> - 使用者必须在遵循 OpenAI 的[使用条款](https://openai.com/policies/terms-of-use)以及**法律法规**的情况下使用，不得用于非法用途。
 > - 本项目仅供个人学习使用，不保证稳定性，且不提供任何技术支持。
+> - 使用者必须在遵循 OpenAI 的[使用条款](https://openai.com/policies/terms-of-use)以及**法律法规**的情况下使用，不得用于非法用途。
 > - 根据[《生成式人工智能服务管理暂行办法》](http://www.cac.gov.cn/2023-07/13/c_1690898327029107.htm)的要求，请勿对中国地区公众提供一切未经备案的生成式人工智能服务。
 
+## 📚 文档
+
+详细文档请访问我们的官方Wiki：[https://docs.newapi.pro/](https://docs.newapi.pro/)
+
+也可访问AI生成的DeepWiki:
+[![Ask DeepWiki](https://deepwiki.com/badge.svg)](https://deepwiki.com/QuantumNous/new-api)
+
 ## ✨ 主要特性
 
-1. 🎨 全新的UI界面（部分界面还待更新）
-2. 🌍 多语言支持（待完善）
-3. 🎨 添加[Midjourney-Proxy(Plus)](https://github.com/novicezk/midjourney-proxy)接口支持，[对接文档](Midjourney.md)
-4. 💰 支持在线充值功能，可在系统设置中设置：
-    - [x] 易支付
-5. 🔍 支持用key查询使用额度：
-    - 配合项目[neko-api-key-tool](https://github.com/Calcium-Ion/neko-api-key-tool)可实现用key查询使用
-6. 📑 分页支持选择每页显示数量
-7. 🔄 兼容原版One API的数据库，可直接使用原版数据库（one-api.db）
-8. 💵 支持模型按次数收费，可在 系统设置-运营设置 中设置
-9. ⚖️ 支持渠道**加权随机**
-10. 📈 数据看板（控制台）
-11. 🔒 可设置令牌能调用的模型
-12. 🤖 支持Telegram授权登录：
-    1. 系统设置-配置登录注册-允许通过Telegram登录
-    2. 对[@Botfather](https://t.me/botfather)输入指令/setdomain
-    3. 选择你的bot，然后输入http(s)://你的网站地址/login
-    4. Telegram Bot 名称是bot username 去掉@后的字符串
-13. 🎵 添加 [Suno API](https://github.com/Suno-API/Suno-API)接口支持，[对接文档](Suno.md)
-14. 🔄 支持Rerank模型，目前兼容Cohere和Jina，可接入Dify，[对接文档](Rerank.md)
-15. ⚡ **[OpenAI Realtime API](https://platform.openai.com/docs/guides/realtime/integration)** - 支持OpenAI的Realtime API，支持Azure渠道
-16. 支持使用路由/chat2link 进入聊天界面
-17. 🧠 支持通过模型名称后缀设置 reasoning effort：
+New API提供了丰富的功能，详细特性请参考[特性说明](https://docs.newapi.pro/wiki/features-introduction)：
+
+1. 🎨 全新的UI界面
+2. 🌍 多语言支持
+3. 💰 支持在线充值功能（易支付）
+4. 🔍 支持用key查询使用额度（配合[neko-api-key-tool](https://github.com/Calcium-Ion/neko-api-key-tool)）
+5. 🔄 兼容原版One API的数据库
+6. 💵 支持模型按次数收费
+7. ⚖️ 支持渠道加权随机
+8. 📈 数据看板（控制台）
+9. 🔒 令牌分组、模型限制
+10. 🤖 支持更多授权登陆方式（LinuxDO,Telegram、OIDC）
+11. 🔄 支持Rerank模型（Cohere和Jina），[接口文档](https://docs.newapi.pro/api/jinaai-rerank)
+12. ⚡ 支持OpenAI Realtime API（包括Azure渠道），[接口文档](https://docs.newapi.pro/api/openai-realtime)
+13. ⚡ 支持Claude Messages 格式，[接口文档](https://docs.newapi.pro/api/anthropic-chat)
+14. 支持使用路由/chat2link进入聊天界面
+15. 🧠 支持通过模型名称后缀设置 reasoning effort：
     1. OpenAI o系列模型
         - 添加后缀 `-high` 设置为 high reasoning effort (例如: `o3-mini-high`)
         - 添加后缀 `-medium` 设置为 medium reasoning effort (例如: `o3-mini-medium`)
         - 添加后缀 `-low` 设置为 low reasoning effort (例如: `o3-mini-low`)
     2. Claude 思考模型
         - 添加后缀 `-thinking` 启用思考模式 (例如: `claude-3-7-sonnet-20250219-thinking`)
-18. 🔄 思考转内容，支持在 `渠道-编辑-渠道额外设置` 中设置 `thinking_to_content` 选项，默认`false`，开启后会将思考内容`reasoning_content`转换为`<think>`标签拼接到内容中返回。
-19. 🔄 模型限流，支持在 `系统设置-速率限制设置` 中设置模型限流，支持设置总请求数限制和成功请求数限制
-20. 💰 缓存计费支持，开启后可以在缓存命中时按照设定的比例计费：
+16. 🔄 思考转内容功能
+17. 🔄 针对用户的模型限流功能
+18. 💰 缓存计费支持，开启后可以在缓存命中时按照设定的比例计费：
     1. 在 `系统设置-运营设置` 中设置 `提示缓存倍率` 选项
     2. 在渠道中设置 `提示缓存倍率`，范围 0-1，例如设置为 0.5 表示缓存命中时按照 50% 计费
     3. 支持的渠道：
         - [x] OpenAI
         - [x] Azure
         - [x] DeepSeek
-        - [ ] Claude
+        - [x] Claude
 
 ## 模型支持
-此版本额外支持以下模型：
+
+此版本支持多种模型，详情请参考[接口文档-中继接口](https://docs.newapi.pro/api)：
+
 1. 第三方模型 **gpts** （gpt-4-gizmo-*）
-2. [Midjourney-Proxy(Plus)](https://github.com/novicezk/midjourney-proxy)接口，[对接文档](Midjourney.md)
-3. 自定义渠道，支持填入完整调用地址
-4. [Suno API](https://github.com/Suno-API/Suno-API) 接口，[对接文档](Suno.md)
-5. Rerank模型，目前支持[Cohere](https://cohere.ai/)和[Jina](https://jina.ai/)，[对接文档](Rerank.md)
-6. Dify
+2. 第三方渠道[Midjourney-Proxy(Plus)](https://github.com/novicezk/midjourney-proxy)接口，[接口文档](https://docs.newapi.pro/api/midjourney-proxy-image)
+3. 第三方渠道[Suno API](https://github.com/Suno-API/Suno-API)接口，[接口文档](https://docs.newapi.pro/api/suno-music)
+4. 自定义渠道，支持填入完整调用地址
+5. Rerank模型（[Cohere](https://cohere.ai/)和[Jina](https://jina.ai/)），[接口文档](https://docs.newapi.pro/api/jinaai-rerank)
+6. Claude Messages 格式，[接口文档](https://docs.newapi.pro/api/anthropic-chat)
+7. Dify，当前仅支持chatflow
 
-您可以在渠道中添加自定义模型gpt-4-gizmo-*，此模型并非OpenAI官方模型，而是第三方模型，使用官方key无法调用。
+## 环境变量配置
 
-## 比原版One API多出的配置
-- `GENERATE_DEFAULT_TOKEN`：是否为新注册用户生成初始令牌，默认为 `false`。
-- `STREAMING_TIMEOUT`：设置流式一次回复的超时时间，默认为 60 秒。
-- `DIFY_DEBUG`：设置 Dify 渠道是否输出工作流和节点信息到客户端，默认为 `true`。
-- `FORCE_STREAM_OPTION`：是否覆盖客户端stream_options参数，请求上游返回流模式usage，默认为 `true`，建议开启，不影响客户端传入stream_options参数返回结果。
-- `GET_MEDIA_TOKEN`：是否统计图片token，默认为 `true`，关闭后将不再在本地计算图片token，可能会导致和上游计费不同，此项覆盖 `GET_MEDIA_TOKEN_NOT_STREAM` 选项作用。
-- `GET_MEDIA_TOKEN_NOT_STREAM`：是否在非流（`stream=false`）情况下统计图片token，默认为 `true`。
-- `UPDATE_TASK`：是否更新异步任务（Midjourney、Suno），默认为 `true`，关闭后将不会更新任务进度。
-- `COHERE_SAFETY_SETTING`：Cohere模型[安全设置](https://docs.cohere.com/docs/safety-modes#overview)，可选值为 `NONE`, `CONTEXTUAL`, `STRICT`，默认为 `NONE`。
-- `GEMINI_VISION_MAX_IMAGE_NUM`：Gemini模型最大图片数量，默认为 `16`，设置为 `-1` 则不限制。
-- `MAX_FILE_DOWNLOAD_MB`: 最大文件下载大小，单位 MB，默认为 `20`。
-- `CRYPTO_SECRET`：加密密钥，用于加密数据库内容。
-- `AZURE_DEFAULT_API_VERSION`：Azure渠道默认API版本，如果渠道设置中未指定API版本，则使用此版本，默认为 `2024-12-01-preview`
-- `NOTIFICATION_LIMIT_DURATION_MINUTE`：通知限制的持续时间（分钟），默认为 `10`。
-- `NOTIFY_LIMIT_COUNT`：用户通知在指定持续时间内的最大数量，默认为 `2`。
+详细配置说明请参考[安装指南-环境变量配置](https://docs.newapi.pro/installation/environment-variables)：
 
-## 已废弃的环境变量
-- ~~`GEMINI_MODEL_MAP`（已废弃）~~：改为到`设置-模型相关设置`中设置
-- ~~`GEMINI_SAFETY_SETTING`（已废弃）~~：改为到`设置-模型相关设置`中设置
+- `GENERATE_DEFAULT_TOKEN`：是否为新注册用户生成初始令牌，默认为 `false`
+- `STREAMING_TIMEOUT`：流式回复超时时间，默认60秒
+- `DIFY_DEBUG`：Dify渠道是否输出工作流和节点信息，默认 `true`
+- `FORCE_STREAM_OPTION`：是否覆盖客户端stream_options参数，默认 `true`
+- `GET_MEDIA_TOKEN`：是否统计图片token，默认 `true`
+- `GET_MEDIA_TOKEN_NOT_STREAM`：非流情况下是否统计图片token，默认 `true`
+- `UPDATE_TASK`：是否更新异步任务（Midjourney、Suno），默认 `true`
+- `COHERE_SAFETY_SETTING`：Cohere模型安全设置，可选值为 `NONE`, `CONTEXTUAL`, `STRICT`，默认 `NONE`
+- `GEMINI_VISION_MAX_IMAGE_NUM`：Gemini模型最大图片数量，默认 `16`
+- `MAX_FILE_DOWNLOAD_MB`: 最大文件下载大小，单位MB，默认 `20`
+- `CRYPTO_SECRET`：加密密钥，用于加密数据库内容
+- `AZURE_DEFAULT_API_VERSION`：Azure渠道默认API版本，默认 `2025-04-01-preview`
+- `NOTIFICATION_LIMIT_DURATION_MINUTE`：通知限制持续时间，默认 `10`分钟
+- `NOTIFY_LIMIT_COUNT`：用户通知在指定持续时间内的最大数量，默认 `2`
+- `ERROR_LOG_ENABLED=true`: 是否记录并显示错误日志，默认`false`
 
 ## 部署
 
+详细部署指南请参考[安装指南-部署方式](https://docs.newapi.pro/installation)：
+
 > [!TIP]
 > 最新版Docker镜像：`calciumion/new-api:latest`  
-> 默认账号root 密码123456
 
-### 多机部署
-- 必须设置环境变量 `SESSION_SECRET`，否则会导致多机部署时登录状态不一致。
-- 如果公用Redis，必须设置 `CRYPTO_SECRET`，否则会导致多机部署时Redis内容无法获取。
+### 多机部署注意事项
+- 必须设置环境变量 `SESSION_SECRET`，否则会导致多机部署时登录状态不一致
+- 如果公用Redis，必须设置 `CRYPTO_SECRET`，否则会导致多机部署时Redis内容无法获取
 
 ### 部署要求
-- 本地数据库（默认）：SQLite（Docker 部署默认使用 SQLite，必须挂载 `/data` 目录到宿主机）
-- 远程数据库：MySQL 版本 >= 5.7.8，PgSQL 版本 >= 9.6
+- 本地数据库（默认）：SQLite（Docker部署必须挂载`/data`目录）
+- 远程数据库：MySQL版本 >= 5.7.8，PgSQL版本 >= 9.6
 
-### 使用宝塔面板Docker功能部署
-安装宝塔面板 (**9.2.0版本**及以上)，前往 [宝塔面板](https://www.bt.cn/new/download.html) 官网，选择正式版的脚本下载安装  
-安装后登录宝塔面板，在菜单栏中点击 Docker ，首次进入会提示安装 Docker 服务，点击立即安装，按提示完成安装  
-安装完成后在应用商店中找到 **New-API** ，点击安装，配置基本选项 即可完成安装  
-[图文教程](BT.md)
+### 部署方式
 
-### 基于 Docker 进行部署
+#### 使用宝塔面板Docker功能部署
+安装宝塔面板（**9.2.0版本**及以上），在应用商店中找到**New-API**安装即可。
+[图文教程](./docs/BT.md)
 
-> [!TIP]
-> 默认管理员账号root 密码123456
-
-### 使用 Docker Compose 部署（推荐）
+#### 使用Docker Compose部署（推荐）
 ```shell
 # 下载项目
 git clone https://github.com/Calcium-Ion/new-api.git
 cd new-api
-# 按需编辑 docker-compose.yml
-# nano docker-compose.yml
-# vim docker-compose.yml
+# 按需编辑docker-compose.yml
 # 启动
 docker-compose up -d
 ```
 
-#### 更新版本
+#### 直接使用Docker镜像
 ```shell
-docker-compose pull
-docker-compose up -d
-```
-
-### 直接使用 Docker 镜像
-```shell
-# 使用 SQLite 的部署命令：
+# 使用SQLite
 docker run --name new-api -d --restart always -p 3000:3000 -e TZ=Asia/Shanghai -v /home/ubuntu/data/new-api:/data calciumion/new-api:latest
 
-# 使用 MySQL 的部署命令，在上面的基础上添加 `-e SQL_DSN="root:123456@tcp(localhost:3306)/oneapi"`，请自行修改数据库连接参数。
-# 例如：
+# 使用MySQL
 docker run --name new-api -d --restart always -p 3000:3000 -e SQL_DSN="root:123456@tcp(localhost:3306)/oneapi" -e TZ=Asia/Shanghai -v /home/ubuntu/data/new-api:/data calciumion/new-api:latest
 ```
 
-#### 更新版本
-```shell
-# 拉取最新镜像
-docker pull calciumion/new-api:latest
-# 停止并删除旧容器
-docker stop new-api
-docker rm new-api
-# 使用相同参数运行新容器
-docker run --name new-api -d --restart always -p 3000:3000 -e TZ=Asia/Shanghai -v /home/ubuntu/data/new-api:/data calciumion/new-api:latest
-```
+## 渠道重试与缓存
+渠道重试功能已经实现，可以在`设置->运营设置->通用设置`设置重试次数，**建议开启缓存**功能。
 
-或者使用 Watchtower 自动更新（不推荐，可能会导致数据库不兼容）：
-```shell
-docker run --rm -v /var/run/docker.sock:/var/run/docker.sock containrrr/watchtower -cR
-```
-
-## 渠道重试
-渠道重试功能已经实现，可以在`设置->运营设置->通用设置`设置重试次数，**建议开启缓存**功能。  
-如果开启了重试功能，重试使用下一个优先级，以此类推。
 ### 缓存设置方法
-1. `REDIS_CONN_STRING`：设置之后将使用 Redis 作为缓存使用。
-    + 例子：`REDIS_CONN_STRING=redis://default:redispw@localhost:49153`
-2. `MEMORY_CACHE_ENABLED`：启用内存缓存（如果设置了`REDIS_CONN_STRING`，则无需手动设置），会导致用户额度的更新存在一定的延迟，可选值为 `true` 和 `false`，未设置则默认为 `false`。
-    + 例子：`MEMORY_CACHE_ENABLED=true`
-### 为什么有的时候没有重试
-这些错误码不会重试：400，504，524
-### 我想让400也重试
-在`渠道->编辑`中，将`状态码复写`改为
-```json
-{
-  "400": "500"
-}
-```
-可以实现400错误转为500错误，从而重试
+1. `REDIS_CONN_STRING`：设置Redis作为缓存
+2. `MEMORY_CACHE_ENABLED`：启用内存缓存（设置了Redis则无需手动设置）
 
-## Midjourney接口设置文档
-[对接文档](Midjourney.md)
+## 接口文档
 
-## Suno接口设置文档
-[对接文档](Suno.md)
+详细接口文档请参考[接口文档](https://docs.newapi.pro/api)：
 
-## 界面截图
-![image](https://github.com/user-attachments/assets/a0dcd349-5df8-4dc8-9acf-ca272b239919)
-
-
-![image](https://github.com/user-attachments/assets/c7d0f7e1-729c-43e2-ac7c-2cb73b0afc8e)
-
-![image](https://github.com/user-attachments/assets/29f81de5-33fc-4fc5-a5ff-f9b54b653c7c)
-
-![image](https://github.com/user-attachments/assets/4fa53e18-d2c5-477a-9b26-b86e44c71e35)
-
-## 交流群
-<img src="https://github.com/user-attachments/assets/9ca0bc82-e057-4230-a28d-9f198fa022e3" width="200">
+- [聊天接口（Chat）](https://docs.newapi.pro/api/openai-chat)
+- [图像接口（Image）](https://docs.newapi.pro/api/openai-image)
+- [重排序接口（Rerank）](https://docs.newapi.pro/api/jinaai-rerank)
+- [实时对话接口（Realtime）](https://docs.newapi.pro/api/openai-realtime)
+- [Claude聊天接口（messages）](https://docs.newapi.pro/api/anthropic-chat)
 
 ## 相关项目
 - [One API](https://github.com/songquanpeng/one-api)：原版项目
 - [Midjourney-Proxy](https://github.com/novicezk/midjourney-proxy)：Midjourney接口支持
-- [chatnio](https://github.com/Deeptrain-Community/chatnio)：下一代 AI 一站式 B/C 端解决方案
+- [chatnio](https://github.com/Deeptrain-Community/chatnio)：下一代AI一站式B/C端解决方案
 - [neko-api-key-tool](https://github.com/Calcium-Ion/neko-api-key-tool)：用key查询使用额度
 
 其他基于New API的项目：
-- [new-api-horizon](https://github.com/Calcium-Ion/new-api-horizon)：New API高性能优化版，专注于高并发优化，并支持Claude格式
-- [VoAPI](https://github.com/VoAPI/VoAPI)：基于New API的前端美化版本，闭源免费
+- [new-api-horizon](https://github.com/Calcium-Ion/new-api-horizon)：New API高性能优化版
+- [VoAPI](https://github.com/VoAPI/VoAPI)：基于New API的前端美化版本
+
+## 帮助支持
+
+如有问题，请参考[帮助支持](https://docs.newapi.pro/support)：
+- [社区交流](https://docs.newapi.pro/support/community-interaction)
+- [反馈问题](https://docs.newapi.pro/support/feedback-issues)
+- [常见问题](https://docs.newapi.pro/support/faq)
 
 ## 🌟 Star History
 

common/constants.go

@@ -1,8 +1,8 @@
 package common
 
 import (
-	"os"
-	"strconv"
+	//"os"
+	//"strconv"
 	"sync"
 	"time"
 
@@ -62,9 +62,13 @@ var EmailDomainWhitelist = []string{
 	"yahoo.com",
 	"foxmail.com",
 }
+var EmailLoginAuthServerList = []string{
+	"smtp.sendcloud.net",
+	"smtp.azurecomm.net",
+}
 
-var DebugEnabled = os.Getenv("DEBUG") == "true"
-var MemoryCacheEnabled = os.Getenv("MEMORY_CACHE_ENABLED") == "true"
+var DebugEnabled bool
+var MemoryCacheEnabled bool
 
 var LogConsumeEnabled = true
 
@@ -77,7 +81,6 @@ var SMTPToken = ""
 
 var GitHubClientId = ""
 var GitHubClientSecret = ""
-
 var LinuxDOClientId = ""
 var LinuxDOClientSecret = ""
 
@@ -104,22 +107,22 @@ var RetryTimes = 0
 
 //var RootUserEmail = ""
 
-var IsMasterNode = os.Getenv("NODE_TYPE") != "slave"
+var IsMasterNode bool
 
-var requestInterval, _ = strconv.Atoi(os.Getenv("POLLING_INTERVAL"))
-var RequestInterval = time.Duration(requestInterval) * time.Second
+var requestInterval int
+var RequestInterval time.Duration
 
-var SyncFrequency = GetEnvOrDefault("SYNC_FREQUENCY", 60) // unit is second
+var SyncFrequency int // unit is second
 
 var BatchUpdateEnabled = false
-var BatchUpdateInterval = GetEnvOrDefault("BATCH_UPDATE_INTERVAL", 5)
+var BatchUpdateInterval int
 
-var RelayTimeout = GetEnvOrDefault("RELAY_TIMEOUT", 0) // unit is second
+var RelayTimeout int // unit is second
 
-var GeminiSafetySetting = GetEnvOrDefaultString("GEMINI_SAFETY_SETTING", "BLOCK_NONE")
+var GeminiSafetySetting string
 
 // https://docs.cohere.com/docs/safety-modes Type; NONE/CONTEXTUAL/STRICT
-var CohereSafetySetting = GetEnvOrDefaultString("COHERE_SAFETY_SETTING", "NONE")
+var CohereSafetySetting string
 
 const (
 	RequestIdKey = "X-Oneapi-Request-Id"
@@ -146,13 +149,13 @@ var (
 // All duration's unit is seconds
 // Shouldn't larger then RateLimitKeyExpirationDuration
 var (
-	GlobalApiRateLimitEnable   = GetEnvOrDefaultBool("GLOBAL_API_RATE_LIMIT_ENABLE", true)
-	GlobalApiRateLimitNum      = GetEnvOrDefault("GLOBAL_API_RATE_LIMIT", 180)
-	GlobalApiRateLimitDuration = int64(GetEnvOrDefault("GLOBAL_API_RATE_LIMIT_DURATION", 180))
+	GlobalApiRateLimitEnable   bool
+	GlobalApiRateLimitNum      int
+	GlobalApiRateLimitDuration int64
 
-	GlobalWebRateLimitEnable   = GetEnvOrDefaultBool("GLOBAL_WEB_RATE_LIMIT_ENABLE", true)
-	GlobalWebRateLimitNum      = GetEnvOrDefault("GLOBAL_WEB_RATE_LIMIT", 60)
-	GlobalWebRateLimitDuration = int64(GetEnvOrDefault("GLOBAL_WEB_RATE_LIMIT_DURATION", 180))
+	GlobalWebRateLimitEnable   bool
+	GlobalWebRateLimitNum      int
+	GlobalWebRateLimitDuration int64
 
 	UploadRateLimitNum            = 10
 	UploadRateLimitDuration int64 = 60
@@ -235,6 +238,9 @@ const (
 	ChannelTypeMokaAI         = 44
 	ChannelTypeVolcEngine     = 45
 	ChannelTypeBaiduV2        = 46
+	ChannelTypeXinference     = 47
+	ChannelTypeXai            = 48
+	ChannelTypeCoze           = 49
 	ChannelTypeDummy          // this one is only for count, do not add any channel after this
 
 )
@@ -287,4 +293,7 @@ var ChannelBaseURLs = []string{
 	"https://api.moka.ai",                       //44
 	"https://ark.cn-beijing.volces.com",         //45
 	"https://qianfan.baidubce.com",              //46
+	"",                                          //47
+	"https://api.x.ai",                          //48
+	"https://api.coze.cn",                       //49
 }

common/custom-event.go

@@ -44,7 +44,7 @@ var fieldReplacer = strings.NewReplacer(
 	"\r", "\\r")
 
 var dataReplacer = strings.NewReplacer(
-	"\n", "\ndata:",
+	"\n", "\n",
 	"\r", "\\r")
 
 type CustomEvent struct {

common/email.go

@@ -5,6 +5,7 @@ import (
 	"encoding/base64"
 	"fmt"
 	"net/smtp"
+	"slices"
 	"strings"
 	"time"
 )
@@ -79,7 +80,7 @@ func SendEmail(subject string, receiver string, content string) error {
 		if err != nil {
 			return err
 		}
-	} else if isOutlookServer(SMTPAccount) || SMTPServer == "smtp.azurecomm.net" {
+	} else if isOutlookServer(SMTPAccount) || slices.Contains(EmailLoginAuthServerList, SMTPServer) {
 		auth = LoginAuth(SMTPAccount, SMTPToken)
 		err = smtp.SendMail(addr, auth, SMTPFrom, to, mail)
 	} else {

common/init.go

@@ -6,6 +6,8 @@ import (
 	"log"
 	"os"
 	"path/filepath"
+	"strconv"
+	"time"
 )
 
 var (
@@ -66,4 +68,31 @@ func LoadEnv() {
 			}
 		}
 	}
+
+	// Initialize variables from constants.go that were using environment variables
+	DebugEnabled = os.Getenv("DEBUG") == "true"
+	MemoryCacheEnabled = os.Getenv("MEMORY_CACHE_ENABLED") == "true"
+	IsMasterNode = os.Getenv("NODE_TYPE") != "slave"
+
+	// Parse requestInterval and set RequestInterval
+	requestInterval, _ = strconv.Atoi(os.Getenv("POLLING_INTERVAL"))
+	RequestInterval = time.Duration(requestInterval) * time.Second
+
+	// Initialize variables with GetEnvOrDefault
+	SyncFrequency = GetEnvOrDefault("SYNC_FREQUENCY", 60)
+	BatchUpdateInterval = GetEnvOrDefault("BATCH_UPDATE_INTERVAL", 5)
+	RelayTimeout = GetEnvOrDefault("RELAY_TIMEOUT", 0)
+
+	// Initialize string variables with GetEnvOrDefaultString
+	GeminiSafetySetting = GetEnvOrDefaultString("GEMINI_SAFETY_SETTING", "BLOCK_NONE")
+	CohereSafetySetting = GetEnvOrDefaultString("COHERE_SAFETY_SETTING", "NONE")
+
+	// Initialize rate limit variables
+	GlobalApiRateLimitEnable = GetEnvOrDefaultBool("GLOBAL_API_RATE_LIMIT_ENABLE", true)
+	GlobalApiRateLimitNum = GetEnvOrDefault("GLOBAL_API_RATE_LIMIT", 180)
+	GlobalApiRateLimitDuration = int64(GetEnvOrDefault("GLOBAL_API_RATE_LIMIT_DURATION", 180))
+
+	GlobalWebRateLimitEnable = GetEnvOrDefaultBool("GLOBAL_WEB_RATE_LIMIT_ENABLE", true)
+	GlobalWebRateLimitNum = GetEnvOrDefault("GLOBAL_WEB_RATE_LIMIT", 60)
+	GlobalWebRateLimitDuration = int64(GetEnvOrDefault("GLOBAL_WEB_RATE_LIMIT_DURATION", 180))
 }

common/json.go

@@ -0,0 +1,18 @@
+package common
+
+import (
+	"bytes"
+	"encoding/json"
+)
+
+func DecodeJson(data []byte, v any) error {
+	return json.NewDecoder(bytes.NewReader(data)).Decode(v)
+}
+
+func DecodeJsonStr(data string, v any) error {
+	return DecodeJson(StringToByteSlice(data), v)
+}
+
+func EncodeJson(v any) ([]byte, error) {
+	return json.Marshal(v)
+}

common/limiter/limiter.go

@@ -0,0 +1,89 @@
+package limiter
+
+import (
+	"context"
+	_ "embed"
+	"fmt"
+	"github.com/go-redis/redis/v8"
+	"one-api/common"
+	"sync"
+)
+
+//go:embed lua/rate_limit.lua
+var rateLimitScript string
+
+type RedisLimiter struct {
+	client         *redis.Client
+	limitScriptSHA string
+}
+
+var (
+	instance *RedisLimiter
+	once     sync.Once
+)
+
+func New(ctx context.Context, r *redis.Client) *RedisLimiter {
+	once.Do(func() {
+		// 预加载脚本
+		limitSHA, err := r.ScriptLoad(ctx, rateLimitScript).Result()
+		if err != nil {
+			common.SysLog(fmt.Sprintf("Failed to load rate limit script: %v", err))
+		}
+		instance = &RedisLimiter{
+			client:         r,
+			limitScriptSHA: limitSHA,
+		}
+	})
+
+	return instance
+}
+
+func (rl *RedisLimiter) Allow(ctx context.Context, key string, opts ...Option) (bool, error) {
+	// 默认配置
+	config := &Config{
+		Capacity:  10,
+		Rate:      1,
+		Requested: 1,
+	}
+
+	// 应用选项模式
+	for _, opt := range opts {
+		opt(config)
+	}
+
+	// 执行限流
+	result, err := rl.client.EvalSha(
+		ctx,
+		rl.limitScriptSHA,
+		[]string{key},
+		config.Requested,
+		config.Rate,
+		config.Capacity,
+	).Int()
+
+	if err != nil {
+		return false, fmt.Errorf("rate limit failed: %w", err)
+	}
+	return result == 1, nil
+}
+
+// Config 配置选项模式
+type Config struct {
+	Capacity  int64
+	Rate      int64
+	Requested int64
+}
+
+type Option func(*Config)
+
+func WithCapacity(c int64) Option {
+	return func(cfg *Config) { cfg.Capacity = c }
+}
+
+func WithRate(r int64) Option {
+	return func(cfg *Config) { cfg.Rate = r }
+}
+
+func WithRequested(n int64) Option {
+	return func(cfg *Config) { cfg.Requested = n }
+}

common/limiter/lua/rate_limit.lua

@@ -0,0 +1,44 @@
+-- 令牌桶限流器
+-- KEYS[1]: 限流器唯一标识
+-- ARGV[1]: 请求令牌数 (通常为1)
+-- ARGV[2]: 令牌生成速率 (每秒)
+-- ARGV[3]: 桶容量
+
+local key = KEYS[1]
+local requested = tonumber(ARGV[1])
+local rate = tonumber(ARGV[2])
+local capacity = tonumber(ARGV[3])
+
+-- 获取当前时间（Redis服务器时间）
+local now = redis.call('TIME')
+local nowInSeconds = tonumber(now[1])
+
+-- 获取桶状态
+local bucket = redis.call('HMGET', key, 'tokens', 'last_time')
+local tokens = tonumber(bucket[1])
+local last_time = tonumber(bucket[2])
+
+-- 初始化桶（首次请求或过期）
+if not tokens or not last_time then
+    tokens = capacity
+    last_time = nowInSeconds
+else
+    -- 计算新增令牌
+    local elapsed = nowInSeconds - last_time
+    local add_tokens = elapsed * rate
+    tokens = math.min(capacity, tokens + add_tokens)
+    last_time = nowInSeconds
+end
+
+-- 判断是否允许请求
+local allowed = false
+if tokens >= requested then
+    tokens = tokens - requested
+    allowed = true
+end
+
+---- 更新桶状态并设置过期时间
+redis.call('HMSET', key, 'tokens', tokens, 'last_time', last_time)
+--redis.call('EXPIRE', key, math.ceil(capacity / rate) + 60) -- 适当延长过期时间
+
+return allowed and 1 or 0

common/redis.go

@@ -92,12 +92,12 @@ func RedisDel(key string) error {
 	return RDB.Del(ctx, key).Err()
 }
 
-func RedisHDelObj(key string) error {
+func RedisDelKey(key string) error {
 	if DebugEnabled {
-		SysLog(fmt.Sprintf("Redis HDEL: key=%s", key))
+		SysLog(fmt.Sprintf("Redis DEL Key: key=%s", key))
 	}
 	ctx := context.Background()
-	return RDB.HDel(ctx, key).Err()
+	return RDB.Del(ctx, key).Err()
 }
 
 func RedisHSetObj(key string, obj interface{}, expiration time.Duration) error {

common/utils.go

@@ -7,7 +7,6 @@ import (
 	"encoding/base64"
 	"encoding/json"
 	"fmt"
-	"github.com/pkg/errors"
 	"html/template"
 	"io"
 	"log"
@@ -22,6 +21,7 @@ import (
 	"time"
 
 	"github.com/google/uuid"
+	"github.com/pkg/errors"
 )
 
 func OpenBrowser(url string) {

constant/azure.go

@@ -0,0 +1,5 @@
+package constant
+
+import "time"
+
+var AzureNoRemoveDotTime = time.Date(2025, time.May, 10, 0, 0, 0, 0, time.UTC).Unix()

constant/env.go

@@ -4,32 +4,42 @@ import (
 	"one-api/common"
 )
 
-var StreamingTimeout = common.GetEnvOrDefault("STREAMING_TIMEOUT", 60)
-var DifyDebug = common.GetEnvOrDefaultBool("DIFY_DEBUG", true)
-
-var MaxFileDownloadMB = common.GetEnvOrDefault("MAX_FILE_DOWNLOAD_MB", 20)
-
-// ForceStreamOption 覆盖请求参数，强制返回usage信息
-var ForceStreamOption = common.GetEnvOrDefaultBool("FORCE_STREAM_OPTION", true)
-
-var GetMediaToken = common.GetEnvOrDefaultBool("GET_MEDIA_TOKEN", true)
-
-var GetMediaTokenNotStream = common.GetEnvOrDefaultBool("GET_MEDIA_TOKEN_NOT_STREAM", true)
-
-var UpdateTask = common.GetEnvOrDefaultBool("UPDATE_TASK", true)
-
-var AzureDefaultAPIVersion = common.GetEnvOrDefaultString("AZURE_DEFAULT_API_VERSION", "2024-12-01-preview")
+var StreamingTimeout int
+var DifyDebug bool
+var MaxFileDownloadMB int
+var ForceStreamOption bool
+var GetMediaToken bool
+var GetMediaTokenNotStream bool
+var UpdateTask bool
+var AzureDefaultAPIVersion string
+var GeminiVisionMaxImageNum int
+var NotifyLimitCount int
+var NotificationLimitDurationMinute int
+var GenerateDefaultToken bool
+var ErrorLogEnabled bool
 
 //var GeminiModelMap = map[string]string{
 //	"gemini-1.0-pro": "v1",
 //}
 
-var GeminiVisionMaxImageNum = common.GetEnvOrDefault("GEMINI_VISION_MAX_IMAGE_NUM", 16)
-
-var NotifyLimitCount = common.GetEnvOrDefault("NOTIFY_LIMIT_COUNT", 2)
-var NotificationLimitDurationMinute = common.GetEnvOrDefault("NOTIFICATION_LIMIT_DURATION_MINUTE", 10)
-
 func InitEnv() {
+	StreamingTimeout = common.GetEnvOrDefault("STREAMING_TIMEOUT", 60)
+	DifyDebug = common.GetEnvOrDefaultBool("DIFY_DEBUG", true)
+	MaxFileDownloadMB = common.GetEnvOrDefault("MAX_FILE_DOWNLOAD_MB", 20)
+	// ForceStreamOption 覆盖请求参数，强制返回usage信息
+	ForceStreamOption = common.GetEnvOrDefaultBool("FORCE_STREAM_OPTION", true)
+	GetMediaToken = common.GetEnvOrDefaultBool("GET_MEDIA_TOKEN", true)
+	GetMediaTokenNotStream = common.GetEnvOrDefaultBool("GET_MEDIA_TOKEN_NOT_STREAM", true)
+	UpdateTask = common.GetEnvOrDefaultBool("UPDATE_TASK", true)
+	AzureDefaultAPIVersion = common.GetEnvOrDefaultString("AZURE_DEFAULT_API_VERSION", "2025-04-01-preview")
+	GeminiVisionMaxImageNum = common.GetEnvOrDefault("GEMINI_VISION_MAX_IMAGE_NUM", 16)
+	NotifyLimitCount = common.GetEnvOrDefault("NOTIFY_LIMIT_COUNT", 2)
+	NotificationLimitDurationMinute = common.GetEnvOrDefault("NOTIFICATION_LIMIT_DURATION_MINUTE", 10)
+	// GenerateDefaultToken 是否生成初始令牌，默认关闭。
+	GenerateDefaultToken = common.GetEnvOrDefaultBool("GENERATE_DEFAULT_TOKEN", false)
+	// 是否启用错误日志
+	ErrorLogEnabled = common.GetEnvOrDefaultBool("ERROR_LOG_ENABLED", false)
+
 	//modelVersionMapStr := strings.TrimSpace(os.Getenv("GEMINI_MODEL_MAP"))
 	//if modelVersionMapStr == "" {
 	//	return
@@ -43,6 +53,3 @@ func InitEnv() {
 	//	}
 	//}
 }
-
-// GenerateDefaultToken 是否生成初始令牌，默认关闭。
-var GenerateDefaultToken = common.GetEnvOrDefaultBool("GENERATE_DEFAULT_TOKEN", false)

constant/setup.go

@@ -0,0 +1,3 @@
+package constant
+
+var Setup = false

constant/user_setting.go

@@ -1,11 +1,12 @@
 package constant
 
 var (
-	UserSettingNotifyType            = "notify_type"             // QuotaWarningType 额度预警类型
-	UserSettingQuotaWarningThreshold = "quota_warning_threshold" // QuotaWarningThreshold 额度预警阈值
-	UserSettingWebhookUrl            = "webhook_url"             // WebhookUrl webhook地址
-	UserSettingWebhookSecret         = "webhook_secret"          // WebhookSecret webhook密钥
-	UserSettingNotificationEmail     = "notification_email"      // NotificationEmail 通知邮箱地址
+	UserSettingNotifyType            = "notify_type"                    // QuotaWarningType 额度预警类型
+	UserSettingQuotaWarningThreshold = "quota_warning_threshold"        // QuotaWarningThreshold 额度预警阈值
+	UserSettingWebhookUrl            = "webhook_url"                    // WebhookUrl webhook地址
+	UserSettingWebhookSecret         = "webhook_secret"                 // WebhookSecret webhook密钥
+	UserSettingNotificationEmail     = "notification_email"             // NotificationEmail 通知邮箱地址
+	UserAcceptUnsetRatioModel        = "accept_unset_model_ratio_model" // AcceptUnsetRatioModel 是否接受未设置价格的模型
 )
 
 var (

controller/channel-billing.go

@@ -108,6 +108,13 @@ type DeepSeekUsageResponse struct {
 	} `json:"balance_infos"`
 }
 
+type OpenRouterCreditResponse struct {
+	Data struct {
+		TotalCredits float64 `json:"total_credits"`
+		TotalUsage   float64 `json:"total_usage"`
+	} `json:"data"`
+}
+
 // GetAuthHeader get auth header
 func GetAuthHeader(token string) http.Header {
 	h := http.Header{}
@@ -281,6 +288,22 @@ func updateChannelAIGC2DBalance(channel *model.Channel) (float64, error) {
 	return response.TotalAvailable, nil
 }
 
+func updateChannelOpenRouterBalance(channel *model.Channel) (float64, error) {
+	url := "https://openrouter.ai/api/v1/credits"
+	body, err := GetResponseBody("GET", url, channel, GetAuthHeader(channel.Key))
+	if err != nil {
+		return 0, err
+	}
+	response := OpenRouterCreditResponse{}
+	err = json.Unmarshal(body, &response)
+	if err != nil {
+		return 0, err
+	}
+	balance := response.Data.TotalCredits - response.Data.TotalUsage
+	channel.UpdateBalance(balance)
+	return balance, nil
+}
+
 func updateChannelBalance(channel *model.Channel) (float64, error) {
 	baseURL := common.ChannelBaseURLs[channel.Type]
 	if channel.GetBaseURL() == "" {
@@ -307,6 +330,8 @@ func updateChannelBalance(channel *model.Channel) (float64, error) {
 		return updateChannelSiliconFlowBalance(channel)
 	case common.ChannelTypeDeepSeek:
 		return updateChannelDeepSeekBalance(channel)
+	case common.ChannelTypeOpenRouter:
+		return updateChannelOpenRouterBalance(channel)
 	default:
 		return 0, errors.New("尚未实现")
 	}

controller/channel-test.go

@@ -103,11 +103,19 @@ func testChannel(channel *model.Channel, testModel string) (err error, openAIErr
 	}
 
 	request := buildTestRequest(testModel)
-	common.SysLog(fmt.Sprintf("testing channel %d with model %s , info %v ", channel.Id, testModel, info))
+	// 创建一个用于日志的 info 副本，移除 ApiKey
+	logInfo := *info
+	logInfo.ApiKey = ""
+	common.SysLog(fmt.Sprintf("testing channel %d with model %s , info %+v ", channel.Id, testModel, logInfo))
+
+	priceData, err := helper.ModelPriceHelper(c, info, 0, int(request.MaxTokens))
+	if err != nil {
+		return err, nil
+	}
 
 	adaptor.Init(info)
 
-	convertedRequest, err := adaptor.ConvertRequest(c, info, request)
+	convertedRequest, err := adaptor.ConvertOpenAIRequest(c, info, request)
 	if err != nil {
 		return err, nil
 	}
@@ -125,7 +133,7 @@ func testChannel(channel *model.Channel, testModel string) (err error, openAIErr
 	if resp != nil {
 		httpResp = resp.(*http.Response)
 		if httpResp.StatusCode != http.StatusOK {
-			err := service.RelayErrorHandler(httpResp)
+			err := service.RelayErrorHandler(httpResp, true)
 			return fmt.Errorf("status code %d: %s", httpResp.StatusCode, err.Error.Message), err
 		}
 	}
@@ -143,10 +151,7 @@ func testChannel(channel *model.Channel, testModel string) (err error, openAIErr
 		return err, nil
 	}
 	info.PromptTokens = usage.PromptTokens
-	priceData, err := helper.ModelPriceHelper(c, info, usage.PromptTokens, int(request.MaxTokens))
-	if err != nil {
-		return err, nil
-	}
+
 	quota := 0
 	if !priceData.UsePrice {
 		quota = usage.PromptTokens + int(math.Round(float64(usage.CompletionTokens)*priceData.CompletionRatio))
@@ -184,17 +189,21 @@ func buildTestRequest(model string) *dto.GeneralOpenAIRequest {
 		return testRequest
 	}
 	// 并非Embedding 模型
-	if strings.HasPrefix(model, "o1") || strings.HasPrefix(model, "o3") {
+	if strings.HasPrefix(model, "o") {
 		testRequest.MaxCompletionTokens = 10
 	} else if strings.Contains(model, "thinking") {
-		testRequest.MaxTokens = 50
+		if !strings.Contains(model, "claude") {
+			testRequest.MaxTokens = 50
+		}
+	} else if strings.Contains(model, "gemini") {
+		testRequest.MaxTokens = 300
 	} else {
 		testRequest.MaxTokens = 10
 	}
-	content, _ := json.Marshal("hi")
+
 	testMessage := dto.Message{
 		Role:    "user",
-		Content: content,
+		Content: "hi",
 	}
 	testRequest.Model = model
 	testRequest.Messages = append(testRequest.Messages, testMessage)
@@ -262,6 +271,13 @@ func testAllChannels(notify bool) error {
 		disableThreshold = 10000000 // a impossible value
 	}
 	gopool.Go(func() {
+		// 使用 defer 确保无论如何都会重置运行状态，防止死锁
+		defer func() {
+			testAllChannelsLock.Lock()
+			testAllChannelsRunning = false
+			testAllChannelsLock.Unlock()
+		}()
+
 		for _, channel := range channels {
 			isChannelEnabled := channel.Status == common.ChannelStatusEnabled
 			tik := time.Now()
@@ -296,9 +312,7 @@ func testAllChannels(notify bool) error {
 			channel.UpdateResponseTime(milliseconds)
 			time.Sleep(common.RequestInterval)
 		}
-		testAllChannelsLock.Lock()
-		testAllChannelsRunning = false
-		testAllChannelsLock.Unlock()
+		
 		if notify {
 			service.NotifyRootUser(dto.NotifyTypeChannelTest, "通道测试完成", "所有通道测试已完成")
 		}

controller/channel.go

@@ -119,6 +119,12 @@ func FetchUpstreamModels(c *gin.Context) {
 		baseURL = channel.GetBaseURL()
 	}
 	url := fmt.Sprintf("%s/v1/models", baseURL)
+	switch channel.Type {
+	case common.ChannelTypeGemini:
+		url = fmt.Sprintf("%s/v1beta/openai/models", baseURL)
+	case common.ChannelTypeAli:
+		url = fmt.Sprintf("%s/compatible-mode/v1/models", baseURL)
+	}
 	body, err := GetResponseBody("GET", url, channel, GetAuthHeader(channel.Key))
 	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
@@ -139,7 +145,11 @@ func FetchUpstreamModels(c *gin.Context) {
 
 	var ids []string
 	for _, model := range result.Data {
-		ids = append(ids, model.ID)
+		id := model.ID
+		if channel.Type == common.ChannelTypeGemini {
+			id = strings.TrimPrefix(id, "models/")
+		}
+		ids = append(ids, id)
 	}
 
 	c.JSON(http.StatusOK, gin.H{
@@ -613,3 +623,44 @@ func BatchSetChannelTag(c *gin.Context) {
 	})
 	return
 }
+
+func GetTagModels(c *gin.Context) {
+	tag := c.Query("tag")
+	if tag == "" {
+		c.JSON(http.StatusBadRequest, gin.H{
+			"success": false,
+			"message": "tag不能为空",
+		})
+		return
+	}
+
+	channels, err := model.GetChannelsByTag(tag, false) // Assuming false for idSort is fine here
+	if err != nil {
+		c.JSON(http.StatusInternalServerError, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+
+	var longestModels string
+	maxLength := 0
+
+	// Find the longest models string among all channels with the given tag
+	for _, channel := range channels {
+		if channel.Models != "" {
+			currentModels := strings.Split(channel.Models, ",")
+			if len(currentModels) > maxLength {
+				maxLength = len(currentModels)
+				longestModels = channel.Models
+			}
+		}
+	}
+
+	c.JSON(http.StatusOK, gin.H{
+		"success": true,
+		"message": "",
+		"data":    longestModels,
+	})
+	return
+}

controller/image.go

@@ -0,0 +1,9 @@
+package controller
+
+import (
+	"github.com/gin-gonic/gin"
+)
+
+func GetImage(c *gin.Context) {
+
+}

controller/log.go

@@ -196,7 +196,7 @@ func DeleteHistoryLogs(c *gin.Context) {
 		})
 		return
 	}
-	count, err := model.DeleteOldLog(targetTimestamp)
+	count, err := model.DeleteOldLog(c.Request.Context(), targetTimestamp, 100)
 	if err != nil {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,

controller/misc.go

@@ -5,9 +5,12 @@ import (
 	"fmt"
 	"net/http"
 	"one-api/common"
+	"one-api/constant"
+	"one-api/middleware"
 	"one-api/model"
 	"one-api/setting"
 	"one-api/setting/operation_setting"
+	"one-api/setting/system_setting"
 	"strings"
 
 	"github.com/gin-gonic/gin"
@@ -22,52 +25,63 @@ func TestStatus(c *gin.Context) {
 		})
 		return
 	}
+	// 获取HTTP统计信息
+	httpStats := middleware.GetStats()
 	c.JSON(http.StatusOK, gin.H{
-		"success": true,
-		"message": "Server is running",
+		"success":    true,
+		"message":    "Server is running",
+		"http_stats": httpStats,
 	})
 	return
 }
 
 func GetStatus(c *gin.Context) {
+
 	c.JSON(http.StatusOK, gin.H{
 		"success": true,
 		"message": "",
 		"data": gin.H{
-			"version":                  common.Version,
-			"start_time":               common.StartTime,
-			"email_verification":       common.EmailVerificationEnabled,
-			"github_oauth":             common.GitHubOAuthEnabled,
-			"github_client_id":         common.GitHubClientId,
-			"linuxdo_oauth":            common.LinuxDOOAuthEnabled,
-			"linuxdo_client_id":        common.LinuxDOClientId,
-			"telegram_oauth":           common.TelegramOAuthEnabled,
-			"telegram_bot_name":        common.TelegramBotName,
-			"system_name":              common.SystemName,
-			"logo":                     common.Logo,
-			"footer_html":              common.Footer,
-			"wechat_qrcode":            common.WeChatAccountQRCodeImageURL,
-			"wechat_login":             common.WeChatAuthEnabled,
-			"server_address":           setting.ServerAddress,
-			"price":                    setting.Price,
-			"min_topup":                setting.MinTopUp,
-			"turnstile_check":          common.TurnstileCheckEnabled,
-			"turnstile_site_key":       common.TurnstileSiteKey,
-			"top_up_link":              common.TopUpLink,
-			"docs_link":                operation_setting.GetGeneralSetting().DocsLink,
-			"quota_per_unit":           common.QuotaPerUnit,
-			"display_in_currency":      common.DisplayInCurrencyEnabled,
-			"enable_batch_update":      common.BatchUpdateEnabled,
-			"enable_drawing":           common.DrawingEnabled,
-			"enable_task":              common.TaskEnabled,
-			"enable_data_export":       common.DataExportEnabled,
-			"data_export_default_time": common.DataExportDefaultTime,
-			"default_collapse_sidebar": common.DefaultCollapseSidebar,
-			"enable_online_topup":      setting.PayAddress != "" && setting.EpayId != "" && setting.EpayKey != "",
-			"mj_notify_enabled":        setting.MjNotifyEnabled,
-			"chats":                    setting.Chats,
-			"demo_site_enabled":        operation_setting.DemoSiteEnabled,
-			"self_use_mode_enabled":    operation_setting.SelfUseModeEnabled,
+			"version":                     common.Version,
+			"start_time":                  common.StartTime,
+			"email_verification":          common.EmailVerificationEnabled,
+			"github_oauth":                common.GitHubOAuthEnabled,
+			"github_client_id":            common.GitHubClientId,
+			"linuxdo_oauth":               common.LinuxDOOAuthEnabled,
+			"linuxdo_client_id":           common.LinuxDOClientId,
+			"telegram_oauth":              common.TelegramOAuthEnabled,
+			"telegram_bot_name":           common.TelegramBotName,
+			"system_name":                 common.SystemName,
+			"logo":                        common.Logo,
+			"footer_html":                 common.Footer,
+			"wechat_qrcode":               common.WeChatAccountQRCodeImageURL,
+			"wechat_login":                common.WeChatAuthEnabled,
+			"server_address":              setting.ServerAddress,
+			"price":                       setting.Price,
+			"min_topup":                   setting.MinTopUp,
+			"turnstile_check":             common.TurnstileCheckEnabled,
+			"turnstile_site_key":          common.TurnstileSiteKey,
+			"top_up_link":                 common.TopUpLink,
+			"docs_link":                   operation_setting.GetGeneralSetting().DocsLink,
+			"quota_per_unit":              common.QuotaPerUnit,
+			"display_in_currency":         common.DisplayInCurrencyEnabled,
+			"enable_batch_update":         common.BatchUpdateEnabled,
+			"enable_drawing":              common.DrawingEnabled,
+			"enable_task":                 common.TaskEnabled,
+			"enable_data_export":          common.DataExportEnabled,
+			"data_export_default_time":    common.DataExportDefaultTime,
+			"default_collapse_sidebar":    common.DefaultCollapseSidebar,
+			"enable_online_topup":         setting.PayAddress != "" && setting.EpayId != "" && setting.EpayKey != "",
+			"mj_notify_enabled":           setting.MjNotifyEnabled,
+			"chats":                       setting.Chats,
+			"demo_site_enabled":           operation_setting.DemoSiteEnabled,
+			"self_use_mode_enabled":       operation_setting.SelfUseModeEnabled,
+			"oidc_enabled":                system_setting.GetOIDCSettings().Enabled,
+			"oidc_client_id":              system_setting.GetOIDCSettings().ClientId,
+			"oidc_authorization_endpoint": system_setting.GetOIDCSettings().AuthorizationEndpoint,
+			"setup":                       constant.Setup,
+			"api_info":                    setting.GetApiInfo(),
+			"announcements":               setting.GetAnnouncements(),
+			"faq":                         setting.GetFAQ(),
 		},
 	})
 	return

controller/oidc.go

@@ -0,0 +1,240 @@
+package controller
+
+import (
+	"encoding/json"
+	"errors"
+	"fmt"
+	"net/http"
+	"net/url"
+	"one-api/common"
+	"one-api/model"
+	"one-api/setting"
+	"one-api/setting/system_setting"
+	"strconv"
+	"strings"
+	"time"
+
+	"github.com/gin-contrib/sessions"
+	"github.com/gin-gonic/gin"
+)
+
+type OidcResponse struct {
+	AccessToken  string `json:"access_token"`
+	IDToken      string `json:"id_token"`
+	RefreshToken string `json:"refresh_token"`
+	TokenType    string `json:"token_type"`
+	ExpiresIn    int    `json:"expires_in"`
+	Scope        string `json:"scope"`
+}
+
+type OidcUser struct {
+	OpenID            string `json:"sub"`
+	Email             string `json:"email"`
+	Name              string `json:"name"`
+	PreferredUsername string `json:"preferred_username"`
+	Picture           string `json:"picture"`
+}
+
+func getOidcUserInfoByCode(code string) (*OidcUser, error) {
+	if code == "" {
+		return nil, errors.New("无效的参数")
+	}
+
+	values := url.Values{}
+	values.Set("client_id", system_setting.GetOIDCSettings().ClientId)
+	values.Set("client_secret", system_setting.GetOIDCSettings().ClientSecret)
+	values.Set("code", code)
+	values.Set("grant_type", "authorization_code")
+	values.Set("redirect_uri", fmt.Sprintf("%s/oauth/oidc", setting.ServerAddress))
+	formData := values.Encode()
+	req, err := http.NewRequest("POST", system_setting.GetOIDCSettings().TokenEndpoint, strings.NewReader(formData))
+	if err != nil {
+		return nil, err
+	}
+	req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
+	req.Header.Set("Accept", "application/json")
+	client := http.Client{
+		Timeout: 5 * time.Second,
+	}
+	res, err := client.Do(req)
+	if err != nil {
+		common.SysLog(err.Error())
+		return nil, errors.New("无法连接至 OIDC 服务器，请稍后重试！")
+	}
+	defer res.Body.Close()
+	var oidcResponse OidcResponse
+	err = json.NewDecoder(res.Body).Decode(&oidcResponse)
+	if err != nil {
+		return nil, err
+	}
+
+	if oidcResponse.AccessToken == "" {
+		common.SysError("OIDC 获取 Token 失败，请检查设置！")
+		return nil, errors.New("OIDC 获取 Token 失败，请检查设置！")
+	}
+
+	req, err = http.NewRequest("GET", system_setting.GetOIDCSettings().UserInfoEndpoint, nil)
+	if err != nil {
+		return nil, err
+	}
+	req.Header.Set("Authorization", "Bearer "+oidcResponse.AccessToken)
+	res2, err := client.Do(req)
+	if err != nil {
+		common.SysLog(err.Error())
+		return nil, errors.New("无法连接至 OIDC 服务器，请稍后重试！")
+	}
+	defer res2.Body.Close()
+	if res2.StatusCode != http.StatusOK {
+		common.SysError("OIDC 获取用户信息失败！请检查设置！")
+		return nil, errors.New("OIDC 获取用户信息失败！请检查设置！")
+	}
+
+	var oidcUser OidcUser
+	err = json.NewDecoder(res2.Body).Decode(&oidcUser)
+	if err != nil {
+		return nil, err
+	}
+	if oidcUser.OpenID == "" || oidcUser.Email == "" {
+		common.SysError("OIDC 获取用户信息为空！请检查设置！")
+		return nil, errors.New("OIDC 获取用户信息为空！请检查设置！")
+	}
+	return &oidcUser, nil
+}
+
+func OidcAuth(c *gin.Context) {
+	session := sessions.Default(c)
+	state := c.Query("state")
+	if state == "" || session.Get("oauth_state") == nil || state != session.Get("oauth_state").(string) {
+		c.JSON(http.StatusForbidden, gin.H{
+			"success": false,
+			"message": "state is empty or not same",
+		})
+		return
+	}
+	username := session.Get("username")
+	if username != nil {
+		OidcBind(c)
+		return
+	}
+	if !system_setting.GetOIDCSettings().Enabled {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": "管理员未开启通过 OIDC 登录以及注册",
+		})
+		return
+	}
+	code := c.Query("code")
+	oidcUser, err := getOidcUserInfoByCode(code)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	user := model.User{
+		OidcId: oidcUser.OpenID,
+	}
+	if model.IsOidcIdAlreadyTaken(user.OidcId) {
+		err := user.FillUserByOidcId()
+		if err != nil {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": err.Error(),
+			})
+			return
+		}
+	} else {
+		if common.RegisterEnabled {
+			user.Email = oidcUser.Email
+			if oidcUser.PreferredUsername != "" {
+				user.Username = oidcUser.PreferredUsername
+			} else {
+				user.Username = "oidc_" + strconv.Itoa(model.GetMaxUserId()+1)
+			}
+			if oidcUser.Name != "" {
+				user.DisplayName = oidcUser.Name
+			} else {
+				user.DisplayName = "OIDC User"
+			}
+			err := user.Insert(0)
+			if err != nil {
+				c.JSON(http.StatusOK, gin.H{
+					"success": false,
+					"message": err.Error(),
+				})
+				return
+			}
+		} else {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": "管理员关闭了新用户注册",
+			})
+			return
+		}
+	}
+
+	if user.Status != common.UserStatusEnabled {
+		c.JSON(http.StatusOK, gin.H{
+			"message": "用户已被封禁",
+			"success": false,
+		})
+		return
+	}
+	setupLogin(&user, c)
+}
+
+func OidcBind(c *gin.Context) {
+	if !system_setting.GetOIDCSettings().Enabled {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": "管理员未开启通过 OIDC 登录以及注册",
+		})
+		return
+	}
+	code := c.Query("code")
+	oidcUser, err := getOidcUserInfoByCode(code)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	user := model.User{
+		OidcId: oidcUser.OpenID,
+	}
+	if model.IsOidcIdAlreadyTaken(user.OidcId) {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": "该 OIDC 账户已被绑定",
+		})
+		return
+	}
+	session := sessions.Default(c)
+	id := session.Get("id")
+	// id := c.GetInt("id")  // critical bug!
+	user.Id = id.(int)
+	err = user.FillUserById()
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	user.OidcId = oidcUser.OpenID
+	err = user.Update(false)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
+	c.JSON(http.StatusOK, gin.H{
+		"success": true,
+		"message": "bind",
+	})
+	return
+}

controller/option.go

@@ -6,6 +6,7 @@ import (
 	"one-api/common"
 	"one-api/model"
 	"one-api/setting"
+	"one-api/setting/system_setting"
 	"strings"
 
 	"github.com/gin-gonic/gin"
@@ -51,6 +52,14 @@ func UpdateOption(c *gin.Context) {
 			})
 			return
 		}
+	case "oidc.enabled":
+		if option.Value == "true" && system_setting.GetOIDCSettings().ClientId == "" {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": "无法启用 OIDC 登录，请先填入 OIDC Client Id 以及 OIDC Client Secret！",
+			})
+			return
+		}
 	case "LinuxDOOAuthEnabled":
 		if option.Value == "true" && common.LinuxDOClientId == "" {
 			c.JSON(http.StatusOK, gin.H{
@@ -81,6 +90,15 @@ func UpdateOption(c *gin.Context) {
 				"success": false,
 				"message": "无法启用 Turnstile 校验，请先填入 Turnstile 校验相关配置信息！",
 			})
+
+			return
+		}
+	case "TelegramOAuthEnabled":
+		if option.Value == "true" && common.TelegramBotToken == "" {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": "无法启用 Telegram OAuth，请先填入 Telegram Bot Token！",
+			})
 			return
 		}
 	case "GroupRatio":
@@ -92,6 +110,42 @@ func UpdateOption(c *gin.Context) {
 			})
 			return
 		}
+	case "ModelRequestRateLimitGroup":
+		err = setting.CheckModelRequestRateLimitGroup(option.Value)
+		if err != nil {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": err.Error(),
+			})
+			return
+		}
+	case "ApiInfo":
+		err = setting.ValidateApiInfo(option.Value)
+		if err != nil {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": err.Error(),
+			})
+			return
+		}
+	case "Announcements":
+		err = setting.ValidateConsoleSettings(option.Value, "Announcements")
+		if err != nil {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": err.Error(),
+			})
+			return
+		}
+	case "FAQ":
+		err = setting.ValidateConsoleSettings(option.Value, "FAQ")
+		if err != nil {
+			c.JSON(http.StatusOK, gin.H{
+				"success": false,
+				"message": err.Error(),
+			})
+			return
+		}
 	}
 	err = model.UpdateOption(option.Key, option.Value)
 	if err != nil {

controller/relay.go

@@ -4,12 +4,11 @@ import (
 	"bytes"
 	"errors"
 	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/gorilla/websocket"
 	"io"
 	"log"
 	"net/http"
 	"one-api/common"
+	constant2 "one-api/constant"
 	"one-api/dto"
 	"one-api/middleware"
 	"one-api/model"
@@ -19,12 +18,15 @@ import (
 	"one-api/relay/helper"
 	"one-api/service"
 	"strings"
+
+	"github.com/gin-gonic/gin"
+	"github.com/gorilla/websocket"
 )
 
 func relayHandler(c *gin.Context, relayMode int) *dto.OpenAIErrorWithStatusCode {
 	var err *dto.OpenAIErrorWithStatusCode
 	switch relayMode {
-	case relayconstant.RelayModeImagesGenerations:
+	case relayconstant.RelayModeImagesGenerations, relayconstant.RelayModeImagesEdits:
 		err = relay.ImageHelper(c)
 	case relayconstant.RelayModeAudioSpeech:
 		fallthrough
@@ -36,9 +38,33 @@ func relayHandler(c *gin.Context, relayMode int) *dto.OpenAIErrorWithStatusCode
 		err = relay.RerankHelper(c, relayMode)
 	case relayconstant.RelayModeEmbeddings:
 		err = relay.EmbeddingHelper(c)
+	case relayconstant.RelayModeResponses:
+		err = relay.ResponsesHelper(c)
+	case relayconstant.RelayModeGemini:
+		err = relay.GeminiHelper(c)
 	default:
 		err = relay.TextHelper(c)
 	}
+
+	if constant2.ErrorLogEnabled && err != nil {
+		// 保存错误日志到mysql中
+		userId := c.GetInt("id")
+		tokenName := c.GetString("token_name")
+		modelName := c.GetString("original_model")
+		tokenId := c.GetInt("token_id")
+		userGroup := c.GetString("group")
+		channelId := c.GetInt("channel_id")
+		other := make(map[string]interface{})
+		other["error_type"] = err.Error.Type
+		other["error_code"] = err.Error.Code
+		other["status_code"] = err.StatusCode
+		other["channel_id"] = channelId
+		other["channel_name"] = c.GetString("channel_name")
+		other["channel_type"] = c.GetInt("channel_type")
+
+		model.RecordErrorLog(c, userId, channelId, modelName, tokenName, err.Error.Message, tokenId, 0, false, userGroup, other)
+	}
+
 	return err
 }
 
@@ -148,6 +174,50 @@ func WssRelay(c *gin.Context) {
 	}
 }
 
+func RelayClaude(c *gin.Context) {
+	//relayMode := constant.Path2RelayMode(c.Request.URL.Path)
+	requestId := c.GetString(common.RequestIdKey)
+	group := c.GetString("group")
+	originalModel := c.GetString("original_model")
+	var claudeErr *dto.ClaudeErrorWithStatusCode
+
+	for i := 0; i <= common.RetryTimes; i++ {
+		channel, err := getChannel(c, group, originalModel, i)
+		if err != nil {
+			common.LogError(c, err.Error())
+			claudeErr = service.ClaudeErrorWrapperLocal(err, "get_channel_failed", http.StatusInternalServerError)
+			break
+		}
+
+		claudeErr = claudeRequest(c, channel)
+
+		if claudeErr == nil {
+			return // 成功处理请求，直接返回
+		}
+
+		openaiErr := service.ClaudeErrorToOpenAIError(claudeErr)
+
+		go processChannelError(c, channel.Id, channel.Type, channel.Name, channel.GetAutoBan(), openaiErr)
+
+		if !shouldRetry(c, openaiErr, common.RetryTimes-i) {
+			break
+		}
+	}
+	useChannel := c.GetStringSlice("use_channel")
+	if len(useChannel) > 1 {
+		retryLogStr := fmt.Sprintf("重试：%s", strings.Trim(strings.Join(strings.Fields(fmt.Sprint(useChannel)), "->"), "[]"))
+		common.LogInfo(c, retryLogStr)
+	}
+
+	if claudeErr != nil {
+		claudeErr.Error.Message = common.MessageWithRequestId(claudeErr.Error.Message, requestId)
+		c.JSON(claudeErr.StatusCode, gin.H{
+			"type":  "error",
+			"error": claudeErr.Error,
+		})
+	}
+}
+
 func relayRequest(c *gin.Context, relayMode int, channel *model.Channel) *dto.OpenAIErrorWithStatusCode {
 	addUsedChannel(c, channel.Id)
 	requestBody, _ := common.GetRequestBody(c)
@@ -162,6 +232,13 @@ func wssRequest(c *gin.Context, ws *websocket.Conn, relayMode int, channel *mode
 	return relay.WssHelper(c, ws)
 }
 
+func claudeRequest(c *gin.Context, channel *model.Channel) *dto.ClaudeErrorWithStatusCode {
+	addUsedChannel(c, channel.Id)
+	requestBody, _ := common.GetRequestBody(c)
+	c.Request.Body = io.NopCloser(bytes.NewBuffer(requestBody))
+	return relay.ClaudeHelper(c)
+}
+
 func addUsedChannel(c *gin.Context, channelId int) {
 	useChannel := c.GetStringSlice("use_channel")
 	useChannel = append(useChannel, fmt.Sprintf("%d", channelId))

controller/setup.go

@@ -0,0 +1,173 @@
+package controller
+
+import (
+	"github.com/gin-gonic/gin"
+	"one-api/common"
+	"one-api/constant"
+	"one-api/model"
+	"one-api/setting/operation_setting"
+	"time"
+)
+
+type Setup struct {
+	Status       bool   `json:"status"`
+	RootInit     bool   `json:"root_init"`
+	DatabaseType string `json:"database_type"`
+}
+
+type SetupRequest struct {
+	Username           string `json:"username"`
+	Password           string `json:"password"`
+	ConfirmPassword    string `json:"confirmPassword"`
+	SelfUseModeEnabled bool   `json:"SelfUseModeEnabled"`
+	DemoSiteEnabled    bool   `json:"DemoSiteEnabled"`
+}
+
+func GetSetup(c *gin.Context) {
+	setup := Setup{
+		Status: constant.Setup,
+	}
+	if constant.Setup {
+		c.JSON(200, gin.H{
+			"success": true,
+			"data":    setup,
+		})
+		return
+	}
+	setup.RootInit = model.RootUserExists()
+	if common.UsingMySQL {
+		setup.DatabaseType = "mysql"
+	}
+	if common.UsingPostgreSQL {
+		setup.DatabaseType = "postgres"
+	}
+	if common.UsingSQLite {
+		setup.DatabaseType = "sqlite"
+	}
+	c.JSON(200, gin.H{
+		"success": true,
+		"data":    setup,
+	})
+}
+
+func PostSetup(c *gin.Context) {
+	// Check if setup is already completed
+	if constant.Setup {
+		c.JSON(400, gin.H{
+			"success": false,
+			"message": "系统已经初始化完成",
+		})
+		return
+	}
+
+	// Check if root user already exists
+	rootExists := model.RootUserExists()
+
+	var req SetupRequest
+	err := c.ShouldBindJSON(&req)
+	if err != nil {
+		c.JSON(400, gin.H{
+			"success": false,
+			"message": "请求参数有误",
+		})
+		return
+	}
+
+	// If root doesn't exist, validate and create admin account
+	if !rootExists {
+		// Validate password
+		if req.Password != req.ConfirmPassword {
+			c.JSON(400, gin.H{
+				"success": false,
+				"message": "两次输入的密码不一致",
+			})
+			return
+		}
+
+		if len(req.Password) < 8 {
+			c.JSON(400, gin.H{
+				"success": false,
+				"message": "密码长度至少为8个字符",
+			})
+			return
+		}
+
+		// Create root user
+		hashedPassword, err := common.Password2Hash(req.Password)
+		if err != nil {
+			c.JSON(500, gin.H{
+				"success": false,
+				"message": "系统错误: " + err.Error(),
+			})
+			return
+		}
+		rootUser := model.User{
+			Username:    req.Username,
+			Password:    hashedPassword,
+			Role:        common.RoleRootUser,
+			Status:      common.UserStatusEnabled,
+			DisplayName: "Root User",
+			AccessToken: nil,
+			Quota:       100000000,
+		}
+		err = model.DB.Create(&rootUser).Error
+		if err != nil {
+			c.JSON(500, gin.H{
+				"success": false,
+				"message": "创建管理员账号失败: " + err.Error(),
+			})
+			return
+		}
+	}
+
+	// Set operation modes
+	operation_setting.SelfUseModeEnabled = req.SelfUseModeEnabled
+	operation_setting.DemoSiteEnabled = req.DemoSiteEnabled
+
+	// Save operation modes to database for persistence
+	err = model.UpdateOption("SelfUseModeEnabled", boolToString(req.SelfUseModeEnabled))
+	if err != nil {
+		c.JSON(500, gin.H{
+			"success": false,
+			"message": "保存自用模式设置失败: " + err.Error(),
+		})
+		return
+	}
+
+	err = model.UpdateOption("DemoSiteEnabled", boolToString(req.DemoSiteEnabled))
+	if err != nil {
+		c.JSON(500, gin.H{
+			"success": false,
+			"message": "保存演示站点模式设置失败: " + err.Error(),
+		})
+		return
+	}
+
+	// Update setup status
+	constant.Setup = true
+
+	setup := model.Setup{
+		Version:       common.Version,
+		InitializedAt: time.Now().Unix(),
+	}
+	err = model.DB.Create(&setup).Error
+	if err != nil {
+		c.JSON(500, gin.H{
+			"success": false,
+			"message": "系统初始化失败: " + err.Error(),
+		})
+		return
+	}
+
+	c.JSON(200, gin.H{
+		"success": true,
+		"message": "系统初始化成功",
+	})
+}
+
+func boolToString(b bool) string {
+	if b {
+		return "true"
+	}
+	return "false"
+}

controller/topup.go

@@ -106,7 +106,7 @@ func RequestEpay(c *gin.Context) {
 		payType = "wxpay"
 	}
 	callBackAddress := service.GetCallbackAddress()
-	returnUrl, _ := url.Parse(setting.ServerAddress + "/log")
+	returnUrl, _ := url.Parse(setting.ServerAddress + "/console/log")
 	notifyUrl, _ := url.Parse(callBackAddress + "/api/user/epay/notify")
 	tradeNo := fmt.Sprintf("%s%d", common.GetRandomString(6), time.Now().Unix())
 	tradeNo = fmt.Sprintf("USR%dNO%s", id, tradeNo)

controller/uptime_kuma.go

@@ -0,0 +1,169 @@
+package controller
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"net/http"
+	"one-api/common"
+	"strings"
+	"time"
+
+	"github.com/gin-gonic/gin"
+	"golang.org/x/sync/errgroup"
+)
+
+type UptimeKumaMonitor struct {
+	ID   int    `json:"id"`
+	Name string `json:"name"`
+	Type string `json:"type"`
+}
+
+type UptimeKumaGroup struct {
+	ID          int                  `json:"id"`
+	Name        string               `json:"name"`
+	Weight      int                  `json:"weight"`
+	MonitorList []UptimeKumaMonitor  `json:"monitorList"`
+}
+
+type UptimeKumaHeartbeat struct {
+	Status int      `json:"status"`
+	Time   string   `json:"time"`
+	Msg    string   `json:"msg"`
+	Ping   *float64 `json:"ping"`
+}
+
+type UptimeKumaStatusResponse struct {
+	PublicGroupList []UptimeKumaGroup `json:"publicGroupList"`
+}
+
+type UptimeKumaHeartbeatResponse struct {
+	HeartbeatList map[string][]UptimeKumaHeartbeat `json:"heartbeatList"`
+	UptimeList    map[string]float64               `json:"uptimeList"`
+}
+
+type MonitorStatus struct {
+	Name   string  `json:"name"`
+	Uptime float64 `json:"uptime"`
+	Status int     `json:"status"`
+}
+
+var (
+	ErrUpstreamNon200 = errors.New("upstream non-200")
+	ErrTimeout        = errors.New("context deadline exceeded")
+)
+
+func getAndDecode(ctx context.Context, client *http.Client, url string, dest interface{}) error {
+	req, err := http.NewRequestWithContext(ctx, http.MethodGet, url, nil)
+	if err != nil {
+		return err
+	}
+
+	resp, err := client.Do(req)
+	if err != nil {
+		if errors.Is(err, context.DeadlineExceeded) || errors.Is(err, context.Canceled) {
+			return ErrTimeout
+		}
+		return err
+	}
+	defer resp.Body.Close()
+
+	if resp.StatusCode != http.StatusOK {
+		return ErrUpstreamNon200
+	}
+
+	return json.NewDecoder(resp.Body).Decode(dest)
+}
+
+func GetUptimeKumaStatus(c *gin.Context) {
+	common.OptionMapRWMutex.RLock()
+	uptimeKumaUrl := common.OptionMap["UptimeKumaUrl"]
+	slug := common.OptionMap["UptimeKumaSlug"]
+	common.OptionMapRWMutex.RUnlock()
+
+	if uptimeKumaUrl == "" || slug == "" {
+		c.JSON(http.StatusOK, gin.H{
+			"success": true,
+			"message": "",
+			"data":    []MonitorStatus{},
+		})
+		return
+	}
+
+	uptimeKumaUrl = strings.TrimSuffix(uptimeKumaUrl, "/")
+
+	ctx, cancel := context.WithTimeout(c.Request.Context(), 10*time.Second)
+	defer cancel()
+
+	client := &http.Client{}
+
+	statusPageUrl := fmt.Sprintf("%s/api/status-page/%s", uptimeKumaUrl, slug)
+	heartbeatUrl := fmt.Sprintf("%s/api/status-page/heartbeat/%s", uptimeKumaUrl, slug)
+
+	var (
+		statusData    UptimeKumaStatusResponse
+		heartbeatData UptimeKumaHeartbeatResponse
+	)
+
+	g, gCtx := errgroup.WithContext(ctx)
+
+	g.Go(func() error {
+		return getAndDecode(gCtx, client, statusPageUrl, &statusData)
+	})
+
+	g.Go(func() error {
+		return getAndDecode(gCtx, client, heartbeatUrl, &heartbeatData)
+	})
+
+	if err := g.Wait(); err != nil {
+		switch err {
+		case ErrUpstreamNon200:
+			c.JSON(http.StatusBadRequest, gin.H{
+				"success": false,
+				"message": "上游接口出现问题",
+			})
+		case ErrTimeout:
+			c.JSON(http.StatusRequestTimeout, gin.H{
+				"success": false,
+				"message": "请求上游接口超时",
+			})
+		default:
+			c.JSON(http.StatusBadRequest, gin.H{
+				"success": false,
+				"message": err.Error(),
+			})
+		}
+		return
+	}
+
+	var monitors []MonitorStatus
+	for _, group := range statusData.PublicGroupList {
+		for _, monitor := range group.MonitorList {
+			monitorStatus := MonitorStatus{
+				Name:   monitor.Name,
+				Uptime: 0.0,
+				Status: 0,
+			}
+
+			uptimeKey := fmt.Sprintf("%d_24", monitor.ID)
+			if uptime, exists := heartbeatData.UptimeList[uptimeKey]; exists {
+				monitorStatus.Uptime = uptime
+			}
+
+			heartbeatKey := fmt.Sprintf("%d", monitor.ID)
+			if heartbeats, exists := heartbeatData.HeartbeatList[heartbeatKey]; exists && len(heartbeats) > 0 {
+				latestHeartbeat := heartbeats[0]
+				monitorStatus.Status = latestHeartbeat.Status
+			}
+
+			monitors = append(monitors, monitorStatus)
+		}
+	}
+
+	c.JSON(http.StatusOK, gin.H{
+		"success": true,
+		"message": "",
+		"data":    monitors,
+	})
+} 

controller/user.go

@@ -592,7 +592,14 @@ func UpdateSelf(c *gin.Context) {
 		user.Password = "" // rollback to what it should be
 		cleanUser.Password = ""
 	}
-	updatePassword := user.Password != ""
+	updatePassword, err := checkUpdatePassword(user.OriginalPassword, user.Password, cleanUser.Id)
+	if err != nil {
+		c.JSON(http.StatusOK, gin.H{
+			"success": false,
+			"message": err.Error(),
+		})
+		return
+	}
 	if err := cleanUser.Update(updatePassword); err != nil {
 		c.JSON(http.StatusOK, gin.H{
 			"success": false,
@@ -608,6 +615,23 @@ func UpdateSelf(c *gin.Context) {
 	return
 }
 
+func checkUpdatePassword(originalPassword string, newPassword string, userId int) (updatePassword bool, err error) {
+	var currentUser *model.User
+	currentUser, err = model.GetUserById(userId, true)
+	if err != nil {
+		return
+	}
+	if !common.ValidatePasswordAndHash(originalPassword, currentUser.Password) {
+		err = fmt.Errorf("原密码错误")
+		return
+	}
+	if newPassword == "" {
+		return
+	}
+	updatePassword = true
+	return
+}
+
 func DeleteUser(c *gin.Context) {
 	id, err := strconv.Atoi(c.Param("id"))
 	if err != nil {
@@ -913,11 +937,12 @@ func TopUp(c *gin.Context) {
 }
 
 type UpdateUserSettingRequest struct {
-	QuotaWarningType      string  `json:"notify_type"`
-	QuotaWarningThreshold float64 `json:"quota_warning_threshold"`
-	WebhookUrl            string  `json:"webhook_url,omitempty"`
-	WebhookSecret         string  `json:"webhook_secret,omitempty"`
-	NotificationEmail     string  `json:"notification_email,omitempty"`
+	QuotaWarningType           string  `json:"notify_type"`
+	QuotaWarningThreshold      float64 `json:"quota_warning_threshold"`
+	WebhookUrl                 string  `json:"webhook_url,omitempty"`
+	WebhookSecret              string  `json:"webhook_secret,omitempty"`
+	NotificationEmail          string  `json:"notification_email,omitempty"`
+	AcceptUnsetModelRatioModel bool    `json:"accept_unset_model_ratio_model"`
 }
 
 func UpdateUserSetting(c *gin.Context) {
@@ -993,6 +1018,7 @@ func UpdateUserSetting(c *gin.Context) {
 	settings := map[string]interface{}{
 		constant.UserSettingNotifyType:            req.QuotaWarningType,
 		constant.UserSettingQuotaWarningThreshold: req.QuotaWarningThreshold,
+		"accept_unset_model_ratio_model":          req.AcceptUnsetModelRatioModel,
 	}
 
 	// 如果是webhook类型,添加webhook相关设置

docker-compose.yml

@@ -15,6 +15,8 @@ services:
       - SQL_DSN=root:123456@tcp(mysql:3306)/new-api  # Point to the mysql service
       - REDIS_CONN_STRING=redis://redis
       - TZ=Asia/Shanghai
+      - ERROR_LOG_ENABLED=true # 是否启用错误日志记录
+    #      - TIKTOKEN_CACHE_DIR=./tiktoken_cache  # 如果需要使用tiktoken_cache，请取消注释
     #      - SESSION_SECRET=random_string  # 多机部署时设置，必须修改这个随机字符串！！！！！！！
     #      - NODE_TYPE=slave  # Uncomment for slave node in multi-node deployment
     #      - SYNC_FREQUENCY=60  # Uncomment if regular database syncing is needed

docs/channel/other_setting.md

@@ -11,7 +11,7 @@
     - 类型为字符串，填写代理地址（例如 socks5 协议的代理地址）
 
 3. thinking_to_content
-   - 用于标识是否将思考内容`reasoning_conetnt`转换为`<think>`标签拼接到内容中返回
+   - 用于标识是否将思考内容`reasoning_content`转换为`<think>`标签拼接到内容中返回
    - 类型为布尔值，设置为 true 时启用思考内容转换
 
 --------------------------------------------------------------
@@ -30,4 +30,4 @@
 
 --------------------------------------------------------------
 
-通过调整上述 JSON 配置中的值，可以灵活控制渠道的额外行为，比如是否进行格式化以及使用特定的网络代理。
+通过调整上述 JSON 配置中的值，可以灵活控制渠道的额外行为，比如是否进行格式化以及使用特定的网络代理。

docs/installation/BT.md

@@ -1,3 +1,3 @@
-密钥为环境变量SESSION_SECRET
-
-![8285bba413e770fe9620f1bf9b40d44e](https://github.com/user-attachments/assets/7a6fc03e-c457-45e4-b8f9-184508fc26b0)
+密钥为环境变量SESSION_SECRET
+
+![8285bba413e770fe9620f1bf9b40d44e](https://github.com/user-attachments/assets/7a6fc03e-c457-45e4-b8f9-184508fc26b0)

dto/claude.go

@@ -0,0 +1,250 @@
+package dto
+
+import (
+	"encoding/json"
+	"one-api/common"
+)
+
+type ClaudeMetadata struct {
+	UserId string `json:"user_id"`
+}
+
+type ClaudeMediaMessage struct {
+	Type         string               `json:"type,omitempty"`
+	Text         *string              `json:"text,omitempty"`
+	Model        string               `json:"model,omitempty"`
+	Source       *ClaudeMessageSource `json:"source,omitempty"`
+	Usage        *ClaudeUsage         `json:"usage,omitempty"`
+	StopReason   *string              `json:"stop_reason,omitempty"`
+	PartialJson  *string              `json:"partial_json,omitempty"`
+	Role         string               `json:"role,omitempty"`
+	Thinking     string               `json:"thinking,omitempty"`
+	Signature    string               `json:"signature,omitempty"`
+	Delta        string               `json:"delta,omitempty"`
+	CacheControl json.RawMessage      `json:"cache_control,omitempty"`
+	// tool_calls
+	Id        string `json:"id,omitempty"`
+	Name      string `json:"name,omitempty"`
+	Input     any    `json:"input,omitempty"`
+	Content   any    `json:"content,omitempty"`
+	ToolUseId string `json:"tool_use_id,omitempty"`
+}
+
+func (c *ClaudeMediaMessage) SetText(s string) {
+	c.Text = &s
+}
+
+func (c *ClaudeMediaMessage) GetText() string {
+	if c.Text == nil {
+		return ""
+	}
+	return *c.Text
+}
+
+func (c *ClaudeMediaMessage) IsStringContent() bool {
+	if c.Content == nil {
+		return false
+	}
+	_, ok := c.Content.(string)
+	if ok {
+		return true
+	}
+	return false
+}
+
+func (c *ClaudeMediaMessage) GetStringContent() string {
+	if c.Content == nil {
+		return ""
+	}
+	switch c.Content.(type) {
+	case string:
+		return c.Content.(string)
+	case []any:
+		var contentStr string
+		for _, contentItem := range c.Content.([]any) {
+			contentMap, ok := contentItem.(map[string]any)
+			if !ok {
+				continue
+			}
+			if contentMap["type"] == ContentTypeText {
+				if subStr, ok := contentMap["text"].(string); ok {
+					contentStr += subStr
+				}
+			}
+		}
+		return contentStr
+	}
+
+	return ""
+}
+
+func (c *ClaudeMediaMessage) GetJsonRowString() string {
+	jsonContent, _ := json.Marshal(c)
+	return string(jsonContent)
+}
+
+func (c *ClaudeMediaMessage) SetContent(content any) {
+	c.Content = content
+}
+
+func (c *ClaudeMediaMessage) ParseMediaContent() []ClaudeMediaMessage {
+	mediaContent, _ := common.Any2Type[[]ClaudeMediaMessage](c.Content)
+	return mediaContent
+}
+
+type ClaudeMessageSource struct {
+	Type      string `json:"type"`
+	MediaType string `json:"media_type,omitempty"`
+	Data      any    `json:"data,omitempty"`
+	Url       string `json:"url,omitempty"`
+}
+
+type ClaudeMessage struct {
+	Role    string `json:"role"`
+	Content any    `json:"content"`
+}
+
+func (c *ClaudeMessage) IsStringContent() bool {
+	if c.Content == nil {
+		return false
+	}
+	_, ok := c.Content.(string)
+	return ok
+}
+
+func (c *ClaudeMessage) GetStringContent() string {
+	if c.Content == nil {
+		return ""
+	}
+	switch c.Content.(type) {
+	case string:
+		return c.Content.(string)
+	case []any:
+		var contentStr string
+		for _, contentItem := range c.Content.([]any) {
+			contentMap, ok := contentItem.(map[string]any)
+			if !ok {
+				continue
+			}
+			if contentMap["type"] == ContentTypeText {
+				if subStr, ok := contentMap["text"].(string); ok {
+					contentStr += subStr
+				}
+			}
+		}
+		return contentStr
+	}
+
+	return ""
+}
+
+func (c *ClaudeMessage) SetStringContent(content string) {
+	c.Content = content
+}
+
+func (c *ClaudeMessage) ParseContent() ([]ClaudeMediaMessage, error) {
+	return common.Any2Type[[]ClaudeMediaMessage](c.Content)
+}
+
+type Tool struct {
+	Name        string                 `json:"name"`
+	Description string                 `json:"description,omitempty"`
+	InputSchema map[string]interface{} `json:"input_schema"`
+}
+
+type InputSchema struct {
+	Type       string `json:"type"`
+	Properties any    `json:"properties,omitempty"`
+	Required   any    `json:"required,omitempty"`
+}
+
+type ClaudeRequest struct {
+	Model             string          `json:"model"`
+	Prompt            string          `json:"prompt,omitempty"`
+	System            any             `json:"system,omitempty"`
+	Messages          []ClaudeMessage `json:"messages,omitempty"`
+	MaxTokens         uint            `json:"max_tokens,omitempty"`
+	MaxTokensToSample uint            `json:"max_tokens_to_sample,omitempty"`
+	StopSequences     []string        `json:"stop_sequences,omitempty"`
+	Temperature       *float64        `json:"temperature,omitempty"`
+	TopP              float64         `json:"top_p,omitempty"`
+	TopK              int             `json:"top_k,omitempty"`
+	//ClaudeMetadata    `json:"metadata,omitempty"`
+	Stream     bool      `json:"stream,omitempty"`
+	Tools      any       `json:"tools,omitempty"`
+	ToolChoice any       `json:"tool_choice,omitempty"`
+	Thinking   *Thinking `json:"thinking,omitempty"`
+}
+
+type Thinking struct {
+	Type         string `json:"type"`
+	BudgetTokens int    `json:"budget_tokens"`
+}
+
+func (c *ClaudeRequest) IsStringSystem() bool {
+	_, ok := c.System.(string)
+	return ok
+}
+
+func (c *ClaudeRequest) GetStringSystem() string {
+	if c.IsStringSystem() {
+		return c.System.(string)
+	}
+	return ""
+}
+
+func (c *ClaudeRequest) SetStringSystem(system string) {
+	c.System = system
+}
+
+func (c *ClaudeRequest) ParseSystem() []ClaudeMediaMessage {
+	mediaContent, _ := common.Any2Type[[]ClaudeMediaMessage](c.System)
+	return mediaContent
+}
+
+type ClaudeError struct {
+	Type    string `json:"type,omitempty"`
+	Message string `json:"message,omitempty"`
+}
+
+type ClaudeErrorWithStatusCode struct {
+	Error      ClaudeError `json:"error"`
+	StatusCode int         `json:"status_code"`
+	LocalError bool
+}
+
+type ClaudeResponse struct {
+	Id           string               `json:"id,omitempty"`
+	Type         string               `json:"type"`
+	Role         string               `json:"role,omitempty"`
+	Content      []ClaudeMediaMessage `json:"content,omitempty"`
+	Completion   string               `json:"completion,omitempty"`
+	StopReason   string               `json:"stop_reason,omitempty"`
+	Model        string               `json:"model,omitempty"`
+	Error        *ClaudeError         `json:"error,omitempty"`
+	Usage        *ClaudeUsage         `json:"usage,omitempty"`
+	Index        *int                 `json:"index,omitempty"`
+	ContentBlock *ClaudeMediaMessage  `json:"content_block,omitempty"`
+	Delta        *ClaudeMediaMessage  `json:"delta,omitempty"`
+	Message      *ClaudeMediaMessage  `json:"message,omitempty"`
+}
+
+// set index
+func (c *ClaudeResponse) SetIndex(i int) {
+	c.Index = &i
+}
+
+// get index
+func (c *ClaudeResponse) GetIndex() int {
+	if c.Index == nil {
+		return 0
+	}
+	return *c.Index
+}
+
+type ClaudeUsage struct {
+	InputTokens              int `json:"input_tokens"`
+	CacheCreationInputTokens int `json:"cache_creation_input_tokens"`
+	CacheReadInputTokens     int `json:"cache_read_input_tokens"`
+	OutputTokens             int `json:"output_tokens"`
+}

dto/dalle.go

@@ -1,14 +1,20 @@
 package dto
 
+import "encoding/json"
+
 type ImageRequest struct {
-	Model          string `json:"model"`
-	Prompt         string `json:"prompt" binding:"required"`
-	N              int    `json:"n,omitempty"`
-	Size           string `json:"size,omitempty"`
-	Quality        string `json:"quality,omitempty"`
-	ResponseFormat string `json:"response_format,omitempty"`
-	Style          string `json:"style,omitempty"`
-	User           string `json:"user,omitempty"`
+	Model          string          `json:"model"`
+	Prompt         string          `json:"prompt" binding:"required"`
+	N              int             `json:"n,omitempty"`
+	Size           string          `json:"size,omitempty"`
+	Quality        string          `json:"quality,omitempty"`
+	ResponseFormat string          `json:"response_format,omitempty"`
+	Style          string          `json:"style,omitempty"`
+	User           string          `json:"user,omitempty"`
+	ExtraFields    json.RawMessage `json:"extra_fields,omitempty"`
+	Background     string          `json:"background,omitempty"`
+	Moderation     string          `json:"moderation,omitempty"`
+	OutputFormat   string          `json:"output_format,omitempty"`
 }
 
 type ImageResponse struct {

dto/openai_request.go

@@ -2,6 +2,7 @@ package dto
 
 import (
 	"encoding/json"
+	"one-api/common"
 	"strings"
 )
 
@@ -36,21 +37,33 @@ type GeneralOpenAIRequest struct {
 	Input               any               `json:"input,omitempty"`
 	Instruction         string            `json:"instruction,omitempty"`
 	Size                string            `json:"size,omitempty"`
-	Functions           any               `json:"functions,omitempty"`
+	Functions           json.RawMessage   `json:"functions,omitempty"`
 	FrequencyPenalty    float64           `json:"frequency_penalty,omitempty"`
 	PresencePenalty     float64           `json:"presence_penalty,omitempty"`
 	ResponseFormat      *ResponseFormat   `json:"response_format,omitempty"`
-	EncodingFormat      any               `json:"encoding_format,omitempty"`
+	EncodingFormat      json.RawMessage   `json:"encoding_format,omitempty"`
 	Seed                float64           `json:"seed,omitempty"`
+	ParallelTooCalls    *bool             `json:"parallel_tool_calls,omitempty"`
 	Tools               []ToolCallRequest `json:"tools,omitempty"`
 	ToolChoice          any               `json:"tool_choice,omitempty"`
 	User                string            `json:"user,omitempty"`
 	LogProbs            bool              `json:"logprobs,omitempty"`
 	TopLogProbs         int               `json:"top_logprobs,omitempty"`
 	Dimensions          int               `json:"dimensions,omitempty"`
-	Modalities          any               `json:"modalities,omitempty"`
-	Audio               any               `json:"audio,omitempty"`
-	ExtraBody           any               `json:"extra_body,omitempty"`
+	Modalities          json.RawMessage   `json:"modalities,omitempty"`
+	Audio               json.RawMessage   `json:"audio,omitempty"`
+	EnableThinking      any               `json:"enable_thinking,omitempty"` // ali
+	ExtraBody           json.RawMessage   `json:"extra_body,omitempty"`
+	WebSearchOptions    *WebSearchOptions `json:"web_search_options,omitempty"`
+	// OpenRouter Params
+	Reasoning json.RawMessage `json:"reasoning,omitempty"`
+}
+
+func (r *GeneralOpenAIRequest) ToMap() map[string]any {
+	result := make(map[string]any)
+	data, _ := common.EncodeJson(r)
+	_ = common.DecodeJson(data, &result)
+	return result
 }
 
 type ToolCallRequest struct {
@@ -70,11 +83,11 @@ type StreamOptions struct {
 	IncludeUsage bool `json:"include_usage,omitempty"`
 }
 
-func (r GeneralOpenAIRequest) GetMaxTokens() int {
+func (r *GeneralOpenAIRequest) GetMaxTokens() int {
 	return int(r.MaxTokens)
 }
 
-func (r GeneralOpenAIRequest) ParseInput() []string {
+func (r *GeneralOpenAIRequest) ParseInput() []string {
 	if r.Input == nil {
 		return nil
 	}
@@ -94,16 +107,16 @@ func (r GeneralOpenAIRequest) ParseInput() []string {
 }
 
 type Message struct {
-	Role                string          `json:"role"`
-	Content             json.RawMessage `json:"content"`
-	Name                *string         `json:"name,omitempty"`
-	Prefix              *bool           `json:"prefix,omitempty"`
-	ReasoningContent    string          `json:"reasoning_content,omitempty"`
-	Reasoning           string          `json:"reasoning,omitempty"`
-	ToolCalls           json.RawMessage `json:"tool_calls,omitempty"`
-	ToolCallId          string          `json:"tool_call_id,omitempty"`
-	parsedContent       []MediaContent
-	parsedStringContent *string
+	Role             string          `json:"role"`
+	Content          any             `json:"content"`
+	Name             *string         `json:"name,omitempty"`
+	Prefix           *bool           `json:"prefix,omitempty"`
+	ReasoningContent string          `json:"reasoning_content,omitempty"`
+	Reasoning        string          `json:"reasoning,omitempty"`
+	ToolCalls        json.RawMessage `json:"tool_calls,omitempty"`
+	ToolCallId       string          `json:"tool_call_id,omitempty"`
+	parsedContent    []MediaContent
+	//parsedStringContent *string
 }
 
 type MediaContent struct {
@@ -111,11 +124,70 @@ type MediaContent struct {
 	Text       string `json:"text,omitempty"`
 	ImageUrl   any    `json:"image_url,omitempty"`
 	InputAudio any    `json:"input_audio,omitempty"`
+	File       any    `json:"file,omitempty"`
+	VideoUrl   any    `json:"video_url,omitempty"`
+	// OpenRouter Params
+	CacheControl json.RawMessage `json:"cache_control,omitempty"`
+}
+
+func (m *MediaContent) GetImageMedia() *MessageImageUrl {
+	if m.ImageUrl != nil {
+		if _, ok := m.ImageUrl.(*MessageImageUrl); ok {
+			return m.ImageUrl.(*MessageImageUrl)
+		}
+		if itemMap, ok := m.ImageUrl.(map[string]any); ok {
+			out := &MessageImageUrl{
+				Url:      common.Interface2String(itemMap["url"]),
+				Detail:   common.Interface2String(itemMap["detail"]),
+				MimeType: common.Interface2String(itemMap["mime_type"]),
+			}
+			return out
+		}
+	}
+	return nil
+}
+
+func (m *MediaContent) GetInputAudio() *MessageInputAudio {
+	if m.InputAudio != nil {
+		if _, ok := m.InputAudio.(*MessageInputAudio); ok {
+			return m.InputAudio.(*MessageInputAudio)
+		}
+		if itemMap, ok := m.InputAudio.(map[string]any); ok {
+			out := &MessageInputAudio{
+				Data:   common.Interface2String(itemMap["data"]),
+				Format: common.Interface2String(itemMap["format"]),
+			}
+			return out
+		}
+	}
+	return nil
+}
+
+func (m *MediaContent) GetFile() *MessageFile {
+	if m.File != nil {
+		if _, ok := m.File.(*MessageFile); ok {
+			return m.File.(*MessageFile)
+		}
+		if itemMap, ok := m.File.(map[string]any); ok {
+			out := &MessageFile{
+				FileName: common.Interface2String(itemMap["file_name"]),
+				FileData: common.Interface2String(itemMap["file_data"]),
+				FileId:   common.Interface2String(itemMap["file_id"]),
+			}
+			return out
+		}
+	}
+	return nil
 }
 
 type MessageImageUrl struct {
-	Url    string `json:"url"`
-	Detail string `json:"detail"`
+	Url      string `json:"url"`
+	Detail   string `json:"detail"`
+	MimeType string
+}
+
+func (m *MessageImageUrl) IsRemoteImage() bool {
+	return strings.HasPrefix(m.Url, "http")
 }
 
 type MessageInputAudio struct {
@@ -123,10 +195,22 @@ type MessageInputAudio struct {
 	Format string `json:"format"`
 }
 
+type MessageFile struct {
+	FileName string `json:"filename,omitempty"`
+	FileData string `json:"file_data,omitempty"`
+	FileId   string `json:"file_id,omitempty"`
+}
+
+type MessageVideoUrl struct {
+	Url string `json:"url"`
+}
+
 const (
 	ContentTypeText       = "text"
 	ContentTypeImageURL   = "image_url"
 	ContentTypeInputAudio = "input_audio"
+	ContentTypeFile       = "file"
+	ContentTypeVideoUrl   = "video_url" // 阿里百炼视频识别
 )
 
 func (m *Message) GetPrefix() bool {
@@ -157,6 +241,186 @@ func (m *Message) SetToolCalls(toolCalls any) {
 }
 
 func (m *Message) StringContent() string {
+	switch m.Content.(type) {
+	case string:
+		return m.Content.(string)
+	case []any:
+		var contentStr string
+		for _, contentItem := range m.Content.([]any) {
+			contentMap, ok := contentItem.(map[string]any)
+			if !ok {
+				continue
+			}
+			if contentMap["type"] == ContentTypeText {
+				if subStr, ok := contentMap["text"].(string); ok {
+					contentStr += subStr
+				}
+			}
+		}
+		return contentStr
+	}
+
+	return ""
+}
+
+func (m *Message) SetNullContent() {
+	m.Content = nil
+	m.parsedContent = nil
+}
+
+func (m *Message) SetStringContent(content string) {
+	m.Content = content
+	m.parsedContent = nil
+}
+
+func (m *Message) SetMediaContent(content []MediaContent) {
+	m.Content = content
+	m.parsedContent = content
+}
+
+func (m *Message) IsStringContent() bool {
+	_, ok := m.Content.(string)
+	if ok {
+		return true
+	}
+	return false
+}
+
+func (m *Message) ParseContent() []MediaContent {
+	if m.Content == nil {
+		return nil
+	}
+	if len(m.parsedContent) > 0 {
+		return m.parsedContent
+	}
+
+	var contentList []MediaContent
+	// 先尝试解析为字符串
+	content, ok := m.Content.(string)
+	if ok {
+		contentList = []MediaContent{{
+			Type: ContentTypeText,
+			Text: content,
+		}}
+		m.parsedContent = contentList
+		return contentList
+	}
+
+	// 尝试解析为数组
+	//var arrayContent []map[string]interface{}
+
+	arrayContent, ok := m.Content.([]any)
+	if !ok {
+		return contentList
+	}
+
+	for _, contentItemAny := range arrayContent {
+		mediaItem, ok := contentItemAny.(MediaContent)
+		if ok {
+			contentList = append(contentList, mediaItem)
+			continue
+		}
+
+		contentItem, ok := contentItemAny.(map[string]any)
+		if !ok {
+			continue
+		}
+		contentType, ok := contentItem["type"].(string)
+		if !ok {
+			continue
+		}
+
+		switch contentType {
+		case ContentTypeText:
+			if text, ok := contentItem["text"].(string); ok {
+				contentList = append(contentList, MediaContent{
+					Type: ContentTypeText,
+					Text: text,
+				})
+			}
+
+		case ContentTypeImageURL:
+			imageUrl := contentItem["image_url"]
+			temp := &MessageImageUrl{
+				Detail: "high",
+			}
+			switch v := imageUrl.(type) {
+			case string:
+				temp.Url = v
+			case map[string]interface{}:
+				url, ok1 := v["url"].(string)
+				detail, ok2 := v["detail"].(string)
+				if ok2 {
+					temp.Detail = detail
+				}
+				if ok1 {
+					temp.Url = url
+				}
+			}
+			contentList = append(contentList, MediaContent{
+				Type:     ContentTypeImageURL,
+				ImageUrl: temp,
+			})
+
+		case ContentTypeInputAudio:
+			if audioData, ok := contentItem["input_audio"].(map[string]interface{}); ok {
+				data, ok1 := audioData["data"].(string)
+				format, ok2 := audioData["format"].(string)
+				if ok1 && ok2 {
+					temp := &MessageInputAudio{
+						Data:   data,
+						Format: format,
+					}
+					contentList = append(contentList, MediaContent{
+						Type:       ContentTypeInputAudio,
+						InputAudio: temp,
+					})
+				}
+			}
+		case ContentTypeFile:
+			if fileData, ok := contentItem["file"].(map[string]interface{}); ok {
+				fileId, ok3 := fileData["file_id"].(string)
+				if ok3 {
+					contentList = append(contentList, MediaContent{
+						Type: ContentTypeFile,
+						File: &MessageFile{
+							FileId: fileId,
+						},
+					})
+				} else {
+					fileName, ok1 := fileData["filename"].(string)
+					fileDataStr, ok2 := fileData["file_data"].(string)
+					if ok1 && ok2 {
+						contentList = append(contentList, MediaContent{
+							Type: ContentTypeFile,
+							File: &MessageFile{
+								FileName: fileName,
+								FileData: fileDataStr,
+							},
+						})
+					}
+				}
+			}
+		case ContentTypeVideoUrl:
+			if videoUrl, ok := contentItem["video_url"].(string); ok {
+				contentList = append(contentList, MediaContent{
+					Type: ContentTypeVideoUrl,
+					VideoUrl: &MessageVideoUrl{
+						Url: videoUrl,
+					},
+				})
+			}
+		}
+	}
+
+	if len(contentList) > 0 {
+		m.parsedContent = contentList
+	}
+	return contentList
+}
+
+// old code
+/*func (m *Message) StringContent() string {
 	if m.parsedStringContent != nil {
 		return *m.parsedStringContent
 	}
@@ -180,6 +444,12 @@ func (m *Message) StringContent() string {
 	return stringContent
 }
 
+func (m *Message) SetNullContent() {
+	m.Content = nil
+	m.parsedStringContent = nil
+	m.parsedContent = nil
+}
+
 func (m *Message) SetStringContent(content string) {
 	jsonContent, _ := json.Marshal(content)
 	m.Content = jsonContent
@@ -244,46 +514,75 @@ func (m *Message) ParseContent() []MediaContent {
 
 			case ContentTypeImageURL:
 				imageUrl := contentItem["image_url"]
+				temp := &MessageImageUrl{
+					Detail: "high",
+				}
 				switch v := imageUrl.(type) {
 				case string:
-					contentList = append(contentList, MediaContent{
-						Type: ContentTypeImageURL,
-						ImageUrl: MessageImageUrl{
-							Url:    v,
-							Detail: "high",
-						},
-					})
+					temp.Url = v
 				case map[string]interface{}:
 					url, ok1 := v["url"].(string)
 					detail, ok2 := v["detail"].(string)
-					if !ok2 {
-						detail = "high"
+					if ok2 {
+						temp.Detail = detail
 					}
 					if ok1 {
-						contentList = append(contentList, MediaContent{
-							Type: ContentTypeImageURL,
-							ImageUrl: MessageImageUrl{
-								Url:    url,
-								Detail: detail,
-							},
-						})
+						temp.Url = url
 					}
 				}
+				contentList = append(contentList, MediaContent{
+					Type:     ContentTypeImageURL,
+					ImageUrl: temp,
+				})
 
 			case ContentTypeInputAudio:
 				if audioData, ok := contentItem["input_audio"].(map[string]interface{}); ok {
 					data, ok1 := audioData["data"].(string)
 					format, ok2 := audioData["format"].(string)
 					if ok1 && ok2 {
+						temp := &MessageInputAudio{
+							Data:   data,
+							Format: format,
+						}
 						contentList = append(contentList, MediaContent{
-							Type: ContentTypeInputAudio,
-							InputAudio: MessageInputAudio{
-								Data:   data,
-								Format: format,
-							},
+							Type:       ContentTypeInputAudio,
+							InputAudio: temp,
 						})
 					}
 				}
+			case ContentTypeFile:
+				if fileData, ok := contentItem["file"].(map[string]interface{}); ok {
+					fileId, ok3 := fileData["file_id"].(string)
+					if ok3 {
+						contentList = append(contentList, MediaContent{
+							Type: ContentTypeFile,
+							File: &MessageFile{
+								FileId: fileId,
+							},
+						})
+					} else {
+						fileName, ok1 := fileData["filename"].(string)
+						fileDataStr, ok2 := fileData["file_data"].(string)
+						if ok1 && ok2 {
+							contentList = append(contentList, MediaContent{
+								Type: ContentTypeFile,
+								File: &MessageFile{
+									FileName: fileName,
+									FileData: fileDataStr,
+								},
+							})
+						}
+					}
+				}
+			case ContentTypeVideoUrl:
+				if videoUrl, ok := contentItem["video_url"].(string); ok {
+					contentList = append(contentList, MediaContent{
+						Type: ContentTypeVideoUrl,
+						VideoUrl: &MessageVideoUrl{
+							Url: videoUrl,
+						},
+					})
+				}
 			}
 		}
 	}
@@ -292,4 +591,55 @@ func (m *Message) ParseContent() []MediaContent {
 		m.parsedContent = contentList
 	}
 	return contentList
+}*/
+
+type WebSearchOptions struct {
+	SearchContextSize string          `json:"search_context_size,omitempty"`
+	UserLocation      json.RawMessage `json:"user_location,omitempty"`
+}
+
+type OpenAIResponsesRequest struct {
+	Model              string               `json:"model"`
+	Input              json.RawMessage      `json:"input,omitempty"`
+	Include            json.RawMessage      `json:"include,omitempty"`
+	Instructions       json.RawMessage      `json:"instructions,omitempty"`
+	MaxOutputTokens    uint                 `json:"max_output_tokens,omitempty"`
+	Metadata           json.RawMessage      `json:"metadata,omitempty"`
+	ParallelToolCalls  bool                 `json:"parallel_tool_calls,omitempty"`
+	PreviousResponseID string               `json:"previous_response_id,omitempty"`
+	Reasoning          *Reasoning           `json:"reasoning,omitempty"`
+	ServiceTier        string               `json:"service_tier,omitempty"`
+	Store              bool                 `json:"store,omitempty"`
+	Stream             bool                 `json:"stream,omitempty"`
+	Temperature        float64              `json:"temperature,omitempty"`
+	Text               json.RawMessage      `json:"text,omitempty"`
+	ToolChoice         json.RawMessage      `json:"tool_choice,omitempty"`
+	Tools              []ResponsesToolsCall `json:"tools,omitempty"`
+	TopP               float64              `json:"top_p,omitempty"`
+	Truncation         string               `json:"truncation,omitempty"`
+	User               string               `json:"user,omitempty"`
+}
+
+type Reasoning struct {
+	Effort  string `json:"effort,omitempty"`
+	Summary string `json:"summary,omitempty"`
+}
+
+type ResponsesToolsCall struct {
+	Type string `json:"type"`
+	// Web Search
+	UserLocation      json.RawMessage `json:"user_location,omitempty"`
+	SearchContextSize string          `json:"search_context_size,omitempty"`
+	// File Search
+	VectorStoreIds []string        `json:"vector_store_ids,omitempty"`
+	MaxNumResults  uint            `json:"max_num_results,omitempty"`
+	Filters        json.RawMessage `json:"filters,omitempty"`
+	// Computer Use
+	DisplayWidth  uint   `json:"display_width,omitempty"`
+	DisplayHeight uint   `json:"display_height,omitempty"`
+	Environment   string `json:"environment,omitempty"`
+	// Function
+	Name        string          `json:"name,omitempty"`
+	Description string          `json:"description,omitempty"`
+	Parameters  json.RawMessage `json:"parameters,omitempty"`
 }

dto/openai_response.go

@@ -1,20 +1,10 @@
 package dto
 
-type TextResponseWithError struct {
-	Id      string                        `json:"id"`
-	Object  string                        `json:"object"`
-	Created int64                         `json:"created"`
-	Choices []OpenAITextResponseChoice    `json:"choices"`
-	Data    []OpenAIEmbeddingResponseItem `json:"data"`
-	Model   string                        `json:"model"`
-	Usage   `json:"usage"`
-	Error   OpenAIError `json:"error"`
-}
+import "encoding/json"
 
 type SimpleResponse struct {
-	Usage   `json:"usage"`
-	Error   OpenAIError                `json:"error"`
-	Choices []OpenAITextResponseChoice `json:"choices"`
+	Usage `json:"usage"`
+	Error *OpenAIError `json:"error"`
 }
 
 type TextResponse struct {
@@ -38,6 +28,7 @@ type OpenAITextResponse struct {
 	Object  string                     `json:"object"`
 	Created int64                      `json:"created"`
 	Choices []OpenAITextResponseChoice `json:"choices"`
+	Error   *OpenAIError               `json:"error,omitempty"`
 	Usage   `json:"usage"`
 }
 
@@ -125,6 +116,20 @@ type ChatCompletionsStreamResponse struct {
 	Usage             *Usage                                `json:"usage"`
 }
 
+func (c *ChatCompletionsStreamResponse) IsToolCall() bool {
+	if len(c.Choices) == 0 {
+		return false
+	}
+	return len(c.Choices[0].Delta.ToolCalls) > 0
+}
+
+func (c *ChatCompletionsStreamResponse) GetFirstToolCall() *ToolCallResponse {
+	if c.IsToolCall() {
+		return &c.Choices[0].Delta.ToolCalls[0]
+	}
+	return nil
+}
+
 func (c *ChatCompletionsStreamResponse) Copy() *ChatCompletionsStreamResponse {
 	choices := make([]ChatCompletionsStreamResponseChoice, len(c.Choices))
 	copy(choices, c.Choices)
@@ -163,10 +168,93 @@ type CompletionsStreamResponse struct {
 }
 
 type Usage struct {
-	PromptTokens           int                `json:"prompt_tokens"`
-	CompletionTokens       int                `json:"completion_tokens"`
-	TotalTokens            int                `json:"total_tokens"`
-	PromptCacheHitTokens   int                `json:"prompt_cache_hit_tokens,omitempty"`
+	PromptTokens         int `json:"prompt_tokens"`
+	CompletionTokens     int `json:"completion_tokens"`
+	TotalTokens          int `json:"total_tokens"`
+	PromptCacheHitTokens int `json:"prompt_cache_hit_tokens,omitempty"`
+
 	PromptTokensDetails    InputTokenDetails  `json:"prompt_tokens_details"`
 	CompletionTokenDetails OutputTokenDetails `json:"completion_tokens_details"`
+	InputTokens            int                `json:"input_tokens"`
+	OutputTokens           int                `json:"output_tokens"`
+	InputTokensDetails     *InputTokenDetails `json:"input_tokens_details"`
+}
+
+type InputTokenDetails struct {
+	CachedTokens         int `json:"cached_tokens"`
+	CachedCreationTokens int `json:"-"`
+	TextTokens           int `json:"text_tokens"`
+	AudioTokens          int `json:"audio_tokens"`
+	ImageTokens          int `json:"image_tokens"`
+}
+
+type OutputTokenDetails struct {
+	TextTokens      int `json:"text_tokens"`
+	AudioTokens     int `json:"audio_tokens"`
+	ReasoningTokens int `json:"reasoning_tokens"`
+}
+
+type OpenAIResponsesResponse struct {
+	ID                 string               `json:"id"`
+	Object             string               `json:"object"`
+	CreatedAt          int                  `json:"created_at"`
+	Status             string               `json:"status"`
+	Error              *OpenAIError         `json:"error,omitempty"`
+	IncompleteDetails  *IncompleteDetails   `json:"incomplete_details,omitempty"`
+	Instructions       string               `json:"instructions"`
+	MaxOutputTokens    int                  `json:"max_output_tokens"`
+	Model              string               `json:"model"`
+	Output             []ResponsesOutput    `json:"output"`
+	ParallelToolCalls  bool                 `json:"parallel_tool_calls"`
+	PreviousResponseID string               `json:"previous_response_id"`
+	Reasoning          *Reasoning           `json:"reasoning"`
+	Store              bool                 `json:"store"`
+	Temperature        float64              `json:"temperature"`
+	ToolChoice         string               `json:"tool_choice"`
+	Tools              []ResponsesToolsCall `json:"tools"`
+	TopP               float64              `json:"top_p"`
+	Truncation         string               `json:"truncation"`
+	Usage              *Usage               `json:"usage"`
+	User               json.RawMessage      `json:"user"`
+	Metadata           json.RawMessage      `json:"metadata"`
+}
+
+type IncompleteDetails struct {
+	Reasoning string `json:"reasoning"`
+}
+
+type ResponsesOutput struct {
+	Type    string                   `json:"type"`
+	ID      string                   `json:"id"`
+	Status  string                   `json:"status"`
+	Role    string                   `json:"role"`
+	Content []ResponsesOutputContent `json:"content"`
+}
+
+type ResponsesOutputContent struct {
+	Type        string        `json:"type"`
+	Text        string        `json:"text"`
+	Annotations []interface{} `json:"annotations"`
+}
+
+const (
+	BuildInToolWebSearchPreview = "web_search_preview"
+	BuildInToolFileSearch       = "file_search"
+)
+
+const (
+	BuildInCallWebSearchCall = "web_search_call"
+)
+
+const (
+	ResponsesOutputTypeItemAdded = "response.output_item.added"
+	ResponsesOutputTypeItemDone  = "response.output_item.done"
+)
+
+// ResponsesStreamResponse 用于处理 /v1/responses 流式响应
+type ResponsesStreamResponse struct {
+	Type     string                   `json:"type"`
+	Response *OpenAIResponsesResponse `json:"response,omitempty"`
+	Delta    string                   `json:"delta,omitempty"`
+	Item     *ResponsesOutput         `json:"item,omitempty"`
 }

dto/realtime.go

@@ -43,18 +43,6 @@ type RealtimeUsage struct {
 	OutputTokenDetails OutputTokenDetails `json:"output_token_details"`
 }
 
-type InputTokenDetails struct {
-	CachedTokens int `json:"cached_tokens"`
-	TextTokens   int `json:"text_tokens"`
-	AudioTokens  int `json:"audio_tokens"`
-	ImageTokens  int `json:"image_tokens"`
-}
-
-type OutputTokenDetails struct {
-	TextTokens  int `json:"text_tokens"`
-	AudioTokens int `json:"audio_tokens"`
-}
-
 type RealtimeSession struct {
 	Modalities              []string                `json:"modalities"`
 	Instructions            string                  `json:"instructions"`

dto/rerank.go

@@ -5,18 +5,29 @@ type RerankRequest struct {
 	Query           string `json:"query"`
 	Model           string `json:"model"`
 	TopN            int    `json:"top_n"`
-	ReturnDocuments bool   `json:"return_documents,omitempty"`
+	ReturnDocuments *bool  `json:"return_documents,omitempty"`
 	MaxChunkPerDoc  int    `json:"max_chunk_per_doc,omitempty"`
 	OverLapTokens   int    `json:"overlap_tokens,omitempty"`
 }
 
-type RerankResponseDocument struct {
+func (r *RerankRequest) GetReturnDocuments() bool {
+	if r.ReturnDocuments == nil {
+		return false
+	}
+	return *r.ReturnDocuments
+}
+
+type RerankResponseResult struct {
 	Document       any     `json:"document,omitempty"`
 	Index          int     `json:"index"`
 	RelevanceScore float64 `json:"relevance_score"`
 }
 
-type RerankResponse struct {
-	Results []RerankResponseDocument `json:"results"`
-	Usage   Usage                    `json:"usage"`
+type RerankDocument struct {
+	Text any `json:"text"`
+}
+
+type RerankResponse struct {
+	Results []RerankResponseResult `json:"results"`
+	Usage   Usage                  `json:"usage"`
 }

go.mod

@@ -24,13 +24,13 @@ require (
 	github.com/gorilla/websocket v1.5.0
 	github.com/joho/godotenv v1.5.1
 	github.com/pkg/errors v0.9.1
-	github.com/pkoukk/tiktoken-go v0.1.7
 	github.com/samber/lo v1.39.0
 	github.com/shirou/gopsutil v3.21.11+incompatible
 	github.com/shopspring/decimal v1.4.0
-	golang.org/x/crypto v0.27.0
+	github.com/tiktoken-go/tokenizer v0.6.2
+	golang.org/x/crypto v0.35.0
 	golang.org/x/image v0.23.0
-	golang.org/x/net v0.28.0
+	golang.org/x/net v0.35.0
 	gorm.io/driver/mysql v1.4.3
 	gorm.io/driver/postgres v1.5.2
 	gorm.io/gorm v1.25.2
@@ -48,7 +48,7 @@ require (
 	github.com/cloudwego/base64x v0.1.4 // indirect
 	github.com/cloudwego/iasm v0.2.0 // indirect
 	github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f // indirect
-	github.com/dlclark/regexp2 v1.11.0 // indirect
+	github.com/dlclark/regexp2 v1.11.5 // indirect
 	github.com/dustin/go-humanize v1.0.1 // indirect
 	github.com/gabriel-vasile/mimetype v1.4.3 // indirect
 	github.com/gin-contrib/sse v0.1.0 // indirect
@@ -84,9 +84,9 @@ require (
 	github.com/yusufpapurcu/wmi v1.2.3 // indirect
 	golang.org/x/arch v0.12.0 // indirect
 	golang.org/x/exp v0.0.0-20240404231335-c0f41cb1a7a0 // indirect
-	golang.org/x/sync v0.10.0 // indirect
-	golang.org/x/sys v0.27.0 // indirect
-	golang.org/x/text v0.21.0 // indirect
+	golang.org/x/sync v0.11.0 // indirect
+	golang.org/x/sys v0.30.0 // indirect
+	golang.org/x/text v0.22.0 // indirect
 	google.golang.org/protobuf v1.34.2 // indirect
 	gopkg.in/yaml.v3 v3.0.1 // indirect
 	modernc.org/libc v1.22.5 // indirect

go.sum

@@ -38,8 +38,8 @@ github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f h1:lO4WD4F/rVNCu3HqELle0jiPLLBs70cWOduZpkS1E78=
 github.com/dgryski/go-rendezvous v0.0.0-20200823014737-9f7001d12a5f/go.mod h1:cuUVRXasLTGF7a8hSLbxyZXjz+1KgoB3wDUb6vlszIc=
-github.com/dlclark/regexp2 v1.11.0 h1:G/nrcoOa7ZXlpoa/91N3X7mM3r8eIlMBBJZvsz/mxKI=
-github.com/dlclark/regexp2 v1.11.0/go.mod h1:DHkYz0B9wPfa6wondMfaivmHpzrQ3v9q8cnmRbL6yW8=
+github.com/dlclark/regexp2 v1.11.5 h1:Q/sSnsKerHeCkc/jSTNq1oCm7KiVgUMZRDUoRu0JQZQ=
+github.com/dlclark/regexp2 v1.11.5/go.mod h1:DHkYz0B9wPfa6wondMfaivmHpzrQ3v9q8cnmRbL6yW8=
 github.com/dustin/go-humanize v1.0.1 h1:GzkhY7T5VNhEkwH0PVJgjz+fX1rhBrR7pRT3mDkpeCY=
 github.com/dustin/go-humanize v1.0.1/go.mod h1:Mu1zIs6XwVuF/gI1OepvI0qD18qycQx+mFykh5fBlto=
 github.com/fsnotify/fsnotify v1.4.9 h1:hsms1Qyu0jgnwNXIxa+/V/PDsU6CfLf6CNO8H7IWoS4=
@@ -167,8 +167,6 @@ github.com/pelletier/go-toml/v2 v2.2.1/go.mod h1:1t835xjRzz80PqgE6HHgN2JOsmgYu/h
 github.com/pkg/diff v0.0.0-20210226163009-20ebb0f2a09e/go.mod h1:pJLUxLENpZxwdsKMEsNbx1VGcRFpLqf3715MtcvvzbA=
 github.com/pkg/errors v0.9.1 h1:FEBLx1zS214owpjy7qsBeixbURkuhQAwrK5UwLGTwt4=
 github.com/pkg/errors v0.9.1/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
-github.com/pkoukk/tiktoken-go v0.1.7 h1:qOBHXX4PHtvIvmOtyg1EeKlwFRiMKAcoMp4Q+bLQDmw=
-github.com/pkoukk/tiktoken-go v0.1.7/go.mod h1:9NiV+i9mJKGj1rYOT+njbv+ZwA/zJxYdewGl6qVatpg=
 github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
 github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
 github.com/remyoudompheng/bigfft v0.0.0-20200410134404-eec4a21b6bb0/go.mod h1:qqbHyh8v60DhA7CoWK5oRCqLrMHRGoxYCSS9EjAz6Eo=
@@ -197,6 +195,8 @@ github.com/stretchr/testify v1.8.1/go.mod h1:w2LPCIKwWwSfY2zedu0+kehJoqGctiVI29o
 github.com/stretchr/testify v1.8.4/go.mod h1:sz/lmYIOXD/1dqDmKjjqLyZ2RngseejIcXlSw2iwfAo=
 github.com/stretchr/testify v1.9.0 h1:HtqpIVDClZ4nwg75+f6Lvsy/wHu+3BoSGCbBAcpTsTg=
 github.com/stretchr/testify v1.9.0/go.mod h1:r2ic/lqez/lEtzL7wO/rwa5dbSLXVDPFyf8C91i36aY=
+github.com/tiktoken-go/tokenizer v0.6.2 h1:t0GN2DvcUZSFWT/62YOgoqb10y7gSXBGs0A+4VCQK+g=
+github.com/tiktoken-go/tokenizer v0.6.2/go.mod h1:6UCYI/DtOallbmL7sSy30p6YQv60qNyU/4aVigPOx6w=
 github.com/tklauser/go-sysconf v0.3.12 h1:0QaGUFOdQaIVdPgfITYzaTegZvdCjmYO52cSFAEVmqU=
 github.com/tklauser/go-sysconf v0.3.12/go.mod h1:Ho14jnntGE1fpdOqQEEaiKRpvIavV0hSfmBq8nJbHYI=
 github.com/tklauser/numcpus v0.6.1 h1:ng9scYS7az0Bk4OZLvrNXNSAO2Pxr1XXRAPyjhIx+Fk=
@@ -217,18 +217,18 @@ golang.org/x/arch v0.0.0-20210923205945-b76863e36670/go.mod h1:5om86z9Hs0C8fWVUu
 golang.org/x/arch v0.12.0 h1:UsYJhbzPYGsT0HbEdmYcqtCv8UNGvnaL561NnIUvaKg=
 golang.org/x/arch v0.12.0/go.mod h1:FEVrYAQjsQXMVJ1nsMoVVXPZg6p2JE2mx8psSWTDQys=
 golang.org/x/crypto v0.0.0-20210711020723-a769d52b0f97/go.mod h1:GvvjBRRGRdwPK5ydBHafDWAxML/pGHZbMvKqRZ5+Abc=
-golang.org/x/crypto v0.27.0 h1:GXm2NjJrPaiv/h1tb2UH8QfgC/hOf/+z0p6PT8o1w7A=
-golang.org/x/crypto v0.27.0/go.mod h1:1Xngt8kV6Dvbssa53Ziq6Eqn0HqbZi5Z6R0ZpwQzt70=
+golang.org/x/crypto v0.35.0 h1:b15kiHdrGCHrP6LvwaQ3c03kgNhhiMgvlhxHQhmg2Xs=
+golang.org/x/crypto v0.35.0/go.mod h1:dy7dXNW32cAb/6/PRuTNsix8T+vJAqvuIy5Bli/x0YQ=
 golang.org/x/exp v0.0.0-20240404231335-c0f41cb1a7a0 h1:985EYyeCOxTpcgOTJpflJUwOeEz0CQOdPt73OzpE9F8=
 golang.org/x/exp v0.0.0-20240404231335-c0f41cb1a7a0/go.mod h1:/lliqkxwWAhPjf5oSOIJup2XcqJaw8RGS6k3TGEc7GI=
 golang.org/x/image v0.23.0 h1:HseQ7c2OpPKTPVzNjG5fwJsOTCiiwS4QdsYi5XU6H68=
 golang.org/x/image v0.23.0/go.mod h1:wJJBTdLfCCf3tiHa1fNxpZmUI4mmoZvwMCPP0ddoNKY=
 golang.org/x/net v0.0.0-20210226172049-e18ecbb05110/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
-golang.org/x/net v0.28.0 h1:a9JDOJc5GMUJ0+UDqmLT86WiEy7iWyIhz8gz8E4e5hE=
-golang.org/x/net v0.28.0/go.mod h1:yqtgsTWOOnlGLG9GFRrK3++bGOUEkNBoHZc8MEDWPNg=
+golang.org/x/net v0.35.0 h1:T5GQRQb2y08kTAByq9L4/bz8cipCdA8FbRTXewonqY8=
+golang.org/x/net v0.35.0/go.mod h1:EglIi67kWsHKlRzzVMUD93VMSWGFOMSZgxFjparz1Qk=
 golang.org/x/sync v0.0.0-20210220032951-036812b2e83c/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
-golang.org/x/sync v0.10.0 h1:3NQrjDixjgGwUOCaF8w2+VYHv0Ve/vGYSbdkTa98gmQ=
-golang.org/x/sync v0.10.0/go.mod h1:Czt+wKu1gCyEFDUtn0jG5QVvpJ6rzVqr5aXyt9drQfk=
+golang.org/x/sync v0.11.0 h1:GGz8+XQP4FvTTrjZPzNKTMFtSXH80RAzG+5ghFPgK9w=
+golang.org/x/sync v0.11.0/go.mod h1:Czt+wKu1gCyEFDUtn0jG5QVvpJ6rzVqr5aXyt9drQfk=
 golang.org/x/sys v0.0.0-20190916202348-b4ddaad3f8a3/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20200116001909-b77594299b42/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
 golang.org/x/sys v0.0.0-20201119102817-f84b799fce68/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
@@ -239,14 +239,14 @@ golang.org/x/sys v0.0.0-20220110181412-a018aaa089fe/go.mod h1:oPkhp1MJrh7nUepCBc
 golang.org/x/sys v0.6.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.8.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.11.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.27.0 h1:wBqf8DvsY9Y/2P8gAfPDEYNuS30J4lPHJxXSb/nJZ+s=
-golang.org/x/sys v0.27.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
+golang.org/x/sys v0.30.0 h1:QjkSwP/36a20jFYWkSue1YwXzLmsV5Gfq7Eiy72C1uc=
+golang.org/x/sys v0.30.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
 golang.org/x/term v0.0.0-20201126162022-7de9c90e9dd1/go.mod h1:bj7SfCRtBDWHUb9snDiAeCFNEtKQo2Wmx5Cou7ajbmo=
 golang.org/x/text v0.3.2/go.mod h1:bEr9sfX3Q8Zfm5fL9x+3itogRgK3+ptLWKqgva+5dAk=
 golang.org/x/text v0.3.3/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
 golang.org/x/text v0.3.6/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
-golang.org/x/text v0.21.0 h1:zyQAAkrwaneQ066sspRyJaG9VNi/YJ1NfzcGB3hZ/qo=
-golang.org/x/text v0.21.0/go.mod h1:4IBbMaMmOPCJ8SecivzSH54+73PCFmPWxNTLm+vZkEQ=
+golang.org/x/text v0.22.0 h1:bofq7m3/HAFvbF51jz3Q9wLg3jkvSPuiZu/pD1XwgtM=
+golang.org/x/text v0.22.0/go.mod h1:YRoo4H8PVmsu+E3Ou7cqLVH8oXWIHVoX0jqUWALQhfY=
 golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
 golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
 google.golang.org/protobuf v1.26.0-rc.1/go.mod h1:jlhhOSvTdKEhbULTjvd4ARK9grFBp09yW+WbY/TyQbw=

main.go

@@ -12,6 +12,7 @@ import (
 	"one-api/model"
 	"one-api/router"
 	"one-api/service"
+	"one-api/setting/operation_setting"
 	"os"
 	"strconv"
 
@@ -33,7 +34,7 @@ var indexPage []byte
 func main() {
 	err := godotenv.Load(".env")
 	if err != nil {
-		common.SysLog("Support for .env file is disabled")
+		common.SysLog("Support for .env file is disabled: " + err.Error())
 	}
 
 	common.LoadEnv()
@@ -51,6 +52,9 @@ func main() {
 	if err != nil {
 		common.FatalLog("failed to initialize database: " + err.Error())
 	}
+
+	model.CheckSetup()
+
 	// Initialize SQL Database
 	err = model.InitLogDB()
 	if err != nil {
@@ -69,10 +73,15 @@ func main() {
 		common.FatalLog("failed to initialize Redis: " + err.Error())
 	}
 
+	// Initialize model settings
+	operation_setting.InitRatioSettings()
 	// Initialize constants
 	constant.InitEnv()
 	// Initialize options
 	model.InitOptionMap()
+
+	service.InitTokenEncoders()
+
 	if common.RedisEnabled {
 		// for compatibility with old versions
 		common.MemoryCacheEnabled = true
@@ -80,9 +89,22 @@ func main() {
 	if common.MemoryCacheEnabled {
 		common.SysLog("memory cache enabled")
 		common.SysError(fmt.Sprintf("sync frequency: %d seconds", common.SyncFrequency))
-		model.InitChannelCache()
-	}
-	if common.MemoryCacheEnabled {
+
+		// Add panic recovery and retry for InitChannelCache
+		func() {
+			defer func() {
+				if r := recover(); r != nil {
+					common.SysError(fmt.Sprintf("InitChannelCache panic: %v, retrying once", r))
+					// Retry once
+					_, fixErr := model.FixAbility()
+					if fixErr != nil {
+						common.SysError(fmt.Sprintf("InitChannelCache failed: %s", fixErr.Error()))
+					}
+				}
+			}()
+			model.InitChannelCache()
+		}()
+
 		go model.SyncOptions(common.SyncFrequency)
 		go model.SyncChannelCache(common.SyncFrequency)
 	}
@@ -126,8 +148,6 @@ func main() {
 		common.SysLog("pprof enabled")
 	}
 
-	service.InitTokenEncoders()
-
 	// Initialize HTTP server
 	server := gin.New()
 	server.Use(gin.CustomRecovery(func(c *gin.Context, err any) {

makefile

@@ -7,7 +7,7 @@ all: build-frontend start-backend
 
 build-frontend:
 	@echo "Building frontend..."
-	@cd $(FRONTEND_DIR) && npm install && DISABLE_ESLINT_PLUGIN='true' VITE_REACT_APP_VERSION=$(cat VERSION) npm run build
+	@cd $(FRONTEND_DIR) && bun install && DISABLE_ESLINT_PLUGIN='true' VITE_REACT_APP_VERSION=$(cat VERSION) bun run build
 
 start-backend:
 	@echo "Starting backend dev server..."

middleware/auth.go

@@ -1,13 +1,14 @@
 package middleware
 
 import (
-	"github.com/gin-contrib/sessions"
-	"github.com/gin-gonic/gin"
 	"net/http"
 	"one-api/common"
 	"one-api/model"
 	"strconv"
 	"strings"
+
+	"github.com/gin-contrib/sessions"
+	"github.com/gin-gonic/gin"
 )
 
 func validUserInfo(username string, role int) bool {
@@ -174,6 +175,26 @@ func TokenAuth() func(c *gin.Context) {
 			}
 			c.Request.Header.Set("Authorization", "Bearer "+key)
 		}
+		// 检查path包含/v1/messages
+		if strings.Contains(c.Request.URL.Path, "/v1/messages") {
+			// 从x-api-key中获取key
+			key := c.Request.Header.Get("x-api-key")
+			if key != "" {
+				c.Request.Header.Set("Authorization", "Bearer "+key)
+			}
+		}
+		// gemini api 从query中获取key
+		if strings.HasPrefix(c.Request.URL.Path, "/v1beta/models/") {
+			skKey := c.Query("key")
+			if skKey != "" {
+				c.Request.Header.Set("Authorization", "Bearer "+skKey)
+			}
+			// 从x-goog-api-key header中获取key
+			xGoogKey := c.Request.Header.Get("x-goog-api-key")
+			if xGoogKey != "" {
+				c.Request.Header.Set("Authorization", "Bearer "+xGoogKey)
+			}
+		}
 		key := c.Request.Header.Get("Authorization")
 		parts := make([]string, 0)
 		key = strings.TrimPrefix(key, "Bearer ")

middleware/distributor.go

@@ -162,7 +162,15 @@ func getModelRequest(c *gin.Context) (*ModelRequest, bool, error) {
 		}
 		c.Set("platform", string(constant.TaskPlatformSuno))
 		c.Set("relay_mode", relayMode)
-	} else if !strings.HasPrefix(c.Request.URL.Path, "/v1/audio/transcriptions") {
+	} else if strings.HasPrefix(c.Request.URL.Path, "/v1beta/models/") {
+		// Gemini API 路径处理: /v1beta/models/gemini-2.0-flash:generateContent
+		relayMode := relayconstant.RelayModeGemini
+		modelName := extractModelNameFromGeminiPath(c.Request.URL.Path)
+		if modelName != "" {
+			modelRequest.Model = modelName
+		}
+		c.Set("relay_mode", relayMode)
+	} else if !strings.HasPrefix(c.Request.URL.Path, "/v1/audio/transcriptions") && !strings.HasPrefix(c.Request.URL.Path, "/v1/images/edits") {
 		err = common.UnmarshalBodyReusable(c, &modelRequest)
 	}
 	if err != nil {
@@ -184,6 +192,8 @@ func getModelRequest(c *gin.Context) (*ModelRequest, bool, error) {
 	}
 	if strings.HasPrefix(c.Request.URL.Path, "/v1/images/generations") {
 		modelRequest.Model = common.GetStringIfEmpty(modelRequest.Model, "dall-e")
+	} else if strings.HasPrefix(c.Request.URL.Path, "/v1/images/edits") {
+		modelRequest.Model = common.GetStringIfEmpty(c.PostForm("model"), "gpt-image-1")
 	}
 	if strings.HasPrefix(c.Request.URL.Path, "/v1/audio") {
 		relayMode := relayconstant.RelayModeAudioSpeech
@@ -211,7 +221,9 @@ func SetupContextForSelectedChannel(c *gin.Context, channel *model.Channel, mode
 	c.Set("channel_id", channel.Id)
 	c.Set("channel_name", channel.Name)
 	c.Set("channel_type", channel.Type)
+	c.Set("channel_create_time", channel.CreatedTime)
 	c.Set("channel_setting", channel.GetSetting())
+	c.Set("param_override", channel.GetParamOverride())
 	if nil != channel.OpenAIOrganization && "" != *channel.OpenAIOrganization {
 		c.Set("channel_organization", *channel.OpenAIOrganization)
 	}
@@ -236,5 +248,35 @@ func SetupContextForSelectedChannel(c *gin.Context, channel *model.Channel, mode
 		c.Set("api_version", channel.Other)
 	case common.ChannelTypeMokaAI:
 		c.Set("api_version", channel.Other)
+	case common.ChannelTypeCoze:
+		c.Set("bot_id", channel.Other)
 	}
 }
+
+// extractModelNameFromGeminiPath 从 Gemini API URL 路径中提取模型名
+// 输入格式: /v1beta/models/gemini-2.0-flash:generateContent
+// 输出: gemini-2.0-flash
+func extractModelNameFromGeminiPath(path string) string {
+	// 查找 "/models/" 的位置
+	modelsPrefix := "/models/"
+	modelsIndex := strings.Index(path, modelsPrefix)
+	if modelsIndex == -1 {
+		return ""
+	}
+
+	// 从 "/models/" 之后开始提取
+	startIndex := modelsIndex + len(modelsPrefix)
+	if startIndex >= len(path) {
+		return ""
+	}
+
+	// 查找 ":" 的位置，模型名在 ":" 之前
+	colonIndex := strings.Index(path[startIndex:], ":")
+	if colonIndex == -1 {
+		// 如果没有找到 ":"，返回从 "/models/" 到路径结尾的部分
+		return path[startIndex:]
+	}
+
+	// 返回模型名部分
+	return path[startIndex : startIndex+colonIndex]
+}

middleware/model-rate-limit.go

@@ -5,6 +5,8 @@ import (
 	"fmt"
 	"net/http"
 	"one-api/common"
+	"one-api/common/limiter"
+	"one-api/constant"
 	"one-api/setting"
 	"strconv"
 	"time"
@@ -78,21 +80,9 @@ func redisRateLimitHandler(duration int64, totalMaxCount, successMaxCount int) g
 		ctx := context.Background()
 		rdb := common.RDB
 
-		// 1. 检查总请求数限制（当totalMaxCount为0时会自动跳过）
-		totalKey := fmt.Sprintf("rateLimit:%s:%s", ModelRequestRateLimitCountMark, userId)
-		allowed, err := checkRedisRateLimit(ctx, rdb, totalKey, totalMaxCount, duration)
-		if err != nil {
-			fmt.Println("检查总请求数限制失败:", err.Error())
-			abortWithOpenAiMessage(c, http.StatusInternalServerError, "rate_limit_check_failed")
-			return
-		}
-		if !allowed {
-			abortWithOpenAiMessage(c, http.StatusTooManyRequests, fmt.Sprintf("您已达到总请求数限制：%d分钟内最多请求%d次，包括失败次数，请检查您的请求是否正确", setting.ModelRequestRateLimitDurationMinutes, totalMaxCount))
-		}
-
-		// 2. 检查成功请求数限制
+		// 1. 检查成功请求数限制
 		successKey := fmt.Sprintf("rateLimit:%s:%s", ModelRequestRateLimitSuccessCountMark, userId)
-		allowed, err = checkRedisRateLimit(ctx, rdb, successKey, successMaxCount, duration)
+		allowed, err := checkRedisRateLimit(ctx, rdb, successKey, successMaxCount, duration)
 		if err != nil {
 			fmt.Println("检查成功请求数限制失败:", err.Error())
 			abortWithOpenAiMessage(c, http.StatusInternalServerError, "rate_limit_check_failed")
@@ -103,8 +93,29 @@ func redisRateLimitHandler(duration int64, totalMaxCount, successMaxCount int) g
 			return
 		}
 
-		// 3. 记录总请求（当totalMaxCount为0时会自动跳过）
-		recordRedisRequest(ctx, rdb, totalKey, totalMaxCount)
+		//2.检查总请求数限制并记录总请求（当totalMaxCount为0时会自动跳过，使用令牌桶限流器
+		if totalMaxCount > 0 {
+			totalKey := fmt.Sprintf("rateLimit:%s", userId)
+			// 初始化
+			tb := limiter.New(ctx, rdb)
+			allowed, err = tb.Allow(
+				ctx,
+				totalKey,
+				limiter.WithCapacity(int64(totalMaxCount)*duration),
+				limiter.WithRate(int64(totalMaxCount)),
+				limiter.WithRequested(duration),
+			)
+
+			if err != nil {
+				fmt.Println("检查总请求数限制失败:", err.Error())
+				abortWithOpenAiMessage(c, http.StatusInternalServerError, "rate_limit_check_failed")
+				return
+			}
+
+			if !allowed {
+				abortWithOpenAiMessage(c, http.StatusTooManyRequests, fmt.Sprintf("您已达到总请求数限制：%d分钟内最多请求%d次，包括失败次数，请检查您的请求是否正确", setting.ModelRequestRateLimitDurationMinutes, totalMaxCount))
+			}
+		}
 
 		// 4. 处理请求
 		c.Next()
@@ -165,6 +176,19 @@ func ModelRequestRateLimit() func(c *gin.Context) {
 		totalMaxCount := setting.ModelRequestRateLimitCount
 		successMaxCount := setting.ModelRequestRateLimitSuccessCount
 
+		// 获取分组
+		group := c.GetString("token_group")
+		if group == "" {
+			group = c.GetString(constant.ContextKeyUserGroup)
+		}
+
+		//获取分组的限流配置
+		groupTotalCount, groupSuccessCount, found := setting.GetGroupRateLimit(group)
+		if found {
+			totalMaxCount = groupTotalCount
+			successMaxCount = groupSuccessCount
+		}
+
 		// 根据存储类型选择并执行限流处理器
 		if common.RedisEnabled {
 			redisRateLimitHandler(duration, totalMaxCount, successMaxCount)(c)

middleware/stats.go

@@ -0,0 +1,41 @@
+package middleware
+
+import (
+	"sync/atomic"
+
+	"github.com/gin-gonic/gin"
+)
+
+// HTTPStats 存储HTTP统计信息
+type HTTPStats struct {
+	activeConnections int64
+}
+
+var globalStats = &HTTPStats{}
+
+// StatsMiddleware 统计中间件
+func StatsMiddleware() gin.HandlerFunc {
+	return func(c *gin.Context) {
+		// 增加活跃连接数
+		atomic.AddInt64(&globalStats.activeConnections, 1)
+		
+		// 确保在请求结束时减少连接数
+		defer func() {
+			atomic.AddInt64(&globalStats.activeConnections, -1)
+		}()
+		
+		c.Next()
+	}
+}
+
+// StatsInfo 统计信息结构
+type StatsInfo struct {
+	ActiveConnections int64 `json:"active_connections"`
+}
+
+// GetStats 获取统计信息
+func GetStats() StatsInfo {
+	return StatsInfo{
+		ActiveConnections: atomic.LoadInt64(&globalStats.activeConnections),
+	}
+} 

model/ability.go

@@ -50,7 +50,7 @@ func getPriority(group string, model string, retry int) (int, error) {
 	err := DB.Model(&Ability{}).
 		Select("DISTINCT(priority)").
 		Where(groupCol+" = ? and model = ? and enabled = "+trueVal, group, model).
-		Order("priority DESC").              // 按优先级降序排序
+		Order("priority DESC"). // 按优先级降序排序
 		Pluck("priority", &priorities).Error // Pluck用于将查询的结果直接扫描到一个切片中
 
 	if err != nil {
@@ -261,12 +261,28 @@ func FixAbility() (int, error) {
 		common.SysError(fmt.Sprintf("Get channel ids from channel table failed: %s", err.Error()))
 		return 0, err
 	}
-	// Delete abilities of channels that are not in channel table
-	err = DB.Where("channel_id NOT IN (?)", channelIds).Delete(&Ability{}).Error
-	if err != nil {
-		common.SysError(fmt.Sprintf("Delete abilities of channels that are not in channel table failed: %s", err.Error()))
-		return 0, err
+
+	// Delete abilities of channels that are not in channel table - in batches to avoid too many placeholders
+	if len(channelIds) > 0 {
+		// Process deletion in chunks to avoid "too many placeholders" error
+		for _, chunk := range lo.Chunk(channelIds, 100) {
+			err = DB.Where("channel_id NOT IN (?)", chunk).Delete(&Ability{}).Error
+			if err != nil {
+				common.SysError(fmt.Sprintf("Delete abilities of channels (batch) that are not in channel table failed: %s", err.Error()))
+				return 0, err
+			}
+		}
+	} else {
+		// If no channels exist, delete all abilities
+		err = DB.Delete(&Ability{}).Error
+		if err != nil {
+			common.SysError(fmt.Sprintf("Delete all abilities failed: %s", err.Error()))
+			return 0, err
+		}
+		common.SysLog("Delete all abilities successfully")
+		return 0, nil
 	}
+
 	common.SysLog(fmt.Sprintf("Delete abilities of channels that are not in channel table successfully, ids: %v", channelIds))
 	count += len(channelIds)
 
@@ -275,17 +291,26 @@ func FixAbility() (int, error) {
 	err = DB.Table("abilities").Distinct("channel_id").Pluck("channel_id", &abilityChannelIds).Error
 	if err != nil {
 		common.SysError(fmt.Sprintf("Get channel ids from abilities table failed: %s", err.Error()))
-		return 0, err
+		return count, err
 	}
+
 	var channels []Channel
 	if len(abilityChannelIds) == 0 {
 		err = DB.Find(&channels).Error
 	} else {
-		err = DB.Where("id NOT IN (?)", abilityChannelIds).Find(&channels).Error
-	}
-	if err != nil {
-		return 0, err
+		// Process query in chunks to avoid "too many placeholders" error
+		err = nil
+		for _, chunk := range lo.Chunk(abilityChannelIds, 100) {
+			var channelsChunk []Channel
+			err = DB.Where("id NOT IN (?)", chunk).Find(&channelsChunk).Error
+			if err != nil {
+				common.SysError(fmt.Sprintf("Find channels not in abilities table failed: %s", err.Error()))
+				return count, err
+			}
+			channels = append(channels, channelsChunk...)
+		}
 	}
+
 	for _, channel := range channels {
 		err := channel.UpdateAbilities(nil)
 		if err != nil {

model/cache.go

@@ -16,6 +16,9 @@ var channelsIDM map[int]*Channel
 var channelSyncLock sync.RWMutex
 
 func InitChannelCache() {
+	if !common.MemoryCacheEnabled {
+		return
+	}
 	newChannelId2channel := make(map[int]*Channel)
 	var channels []*Channel
 	DB.Where("status = ?", common.ChannelStatusEnabled).Find(&channels)
@@ -84,9 +87,11 @@ func CacheGetRandomSatisfiedChannel(group string, model string, retry int) (*Cha
 	if !common.MemoryCacheEnabled {
 		return GetRandomSatisfiedChannel(group, model, retry)
 	}
+
 	channelSyncLock.RLock()
-	defer channelSyncLock.RUnlock()
 	channels := group2model2channels[group][model]
+	channelSyncLock.RUnlock()
+
 	if len(channels) == 0 {
 		return nil, errors.New("channel not found")
 	}

model/channel.go

@@ -36,6 +36,7 @@ type Channel struct {
 	OtherInfo         string  `json:"other_info"`
 	Tag               *string `json:"tag" gorm:"index"`
 	Setting           *string `json:"setting" gorm:"type:text"`
+	ParamOverride     *string `json:"param_override" gorm:"type:text"`
 }
 
 func (channel *Channel) GetModels() []string {
@@ -45,6 +46,17 @@ func (channel *Channel) GetModels() []string {
 	return strings.Split(strings.Trim(channel.Models, ","), ",")
 }
 
+func (channel *Channel) GetGroups() []string {
+	if channel.Group == "" {
+		return []string{}
+	}
+	groups := strings.Split(strings.Trim(channel.Group, ","), ",")
+	for i, group := range groups {
+		groups[i] = strings.TrimSpace(group)
+	}
+	return groups
+}
+
 func (channel *Channel) GetOtherInfo() map[string]interface{} {
 	otherInfo := make(map[string]interface{})
 	if channel.OtherInfo != "" {
@@ -118,10 +130,15 @@ func SearchChannels(keyword string, group string, model string, idSort bool) ([]
 
 	// 如果是 PostgreSQL，使用双引号
 	if common.UsingPostgreSQL {
-		keyCol = `"key"`
 		modelsCol = `"models"`
 	}
 
+	baseURLCol := "`base_url`"
+	// 如果是 PostgreSQL，使用双引号
+	if common.UsingPostgreSQL {
+		baseURLCol = `"base_url"`
+	}
+
 	order := "priority desc"
 	if idSort {
 		order = "id desc"
@@ -141,11 +158,11 @@ func SearchChannels(keyword string, group string, model string, idSort bool) ([]
 			// sqlite, PostgreSQL
 			groupCondition = `(',' || ` + groupCol + ` || ',') LIKE ?`
 		}
-		whereClause = "(id = ? OR name LIKE ? OR " + keyCol + " = ?) AND " + modelsCol + ` LIKE ? AND ` + groupCondition
-		args = append(args, common.String2Int(keyword), "%"+keyword+"%", keyword, "%"+model+"%", "%,"+group+",%")
+		whereClause = "(id = ? OR name LIKE ? OR " + keyCol + " = ? OR " + baseURLCol + " LIKE ?) AND " + modelsCol + ` LIKE ? AND ` + groupCondition
+		args = append(args, common.String2Int(keyword), "%"+keyword+"%", keyword, "%"+keyword+"%", "%"+model+"%", "%,"+group+",%")
 	} else {
-		whereClause = "(id = ? OR name LIKE ? OR " + keyCol + " = ?) AND " + modelsCol + " LIKE ?"
-		args = append(args, common.String2Int(keyword), "%"+keyword+"%", keyword, "%"+model+"%")
+		whereClause = "(id = ? OR name LIKE ? OR " + keyCol + " = ? OR " + baseURLCol + " LIKE ?) AND " + modelsCol + " LIKE ?"
+		args = append(args, common.String2Int(keyword), "%"+keyword+"%", keyword, "%"+keyword+"%", "%"+model+"%")
 	}
 
 	// 执行查询
@@ -449,6 +466,12 @@ func SearchTags(keyword string, group string, model string, idSort bool) ([]*str
 		modelsCol = `"models"`
 	}
 
+	baseURLCol := "`base_url`"
+	// 如果是 PostgreSQL，使用双引号
+	if common.UsingPostgreSQL {
+		baseURLCol = `"base_url"`
+	}
+
 	order := "priority desc"
 	if idSort {
 		order = "id desc"
@@ -468,11 +491,11 @@ func SearchTags(keyword string, group string, model string, idSort bool) ([]*str
 			// sqlite, PostgreSQL
 			groupCondition = `(',' || ` + groupCol + ` || ',') LIKE ?`
 		}
-		whereClause = "(id = ? OR name LIKE ? OR " + keyCol + " = ?) AND " + modelsCol + ` LIKE ? AND ` + groupCondition
-		args = append(args, common.String2Int(keyword), "%"+keyword+"%", keyword, "%"+model+"%", "%,"+group+",%")
+		whereClause = "(id = ? OR name LIKE ? OR " + keyCol + " = ? OR " + baseURLCol + " LIKE ?) AND " + modelsCol + ` LIKE ? AND ` + groupCondition
+		args = append(args, common.String2Int(keyword), "%"+keyword+"%", keyword, "%"+keyword+"%", "%"+model+"%", "%,"+group+",%")
 	} else {
-		whereClause = "(id = ? OR name LIKE ? OR " + keyCol + " = ?) AND " + modelsCol + " LIKE ?"
-		args = append(args, common.String2Int(keyword), "%"+keyword+"%", keyword, "%"+model+"%")
+		whereClause = "(id = ? OR name LIKE ? OR " + keyCol + " = ? OR " + baseURLCol + " LIKE ?) AND " + modelsCol + " LIKE ?"
+		args = append(args, common.String2Int(keyword), "%"+keyword+"%", keyword, "%"+keyword+"%", "%"+model+"%")
 	}
 
 	subQuery := baseQuery.Where(whereClause, args...).
@@ -511,6 +534,17 @@ func (channel *Channel) SetSetting(setting map[string]interface{}) {
 	channel.Setting = common.GetPointer[string](string(settingBytes))
 }
 
+func (channel *Channel) GetParamOverride() map[string]interface{} {
+	paramOverride := make(map[string]interface{})
+	if channel.ParamOverride != nil && *channel.ParamOverride != "" {
+		err := json.Unmarshal([]byte(*channel.ParamOverride), &paramOverride)
+		if err != nil {
+			common.SysError("failed to unmarshal param override: " + err.Error())
+		}
+	}
+	return paramOverride
+}
+
 func GetChannelsByIds(ids []int) ([]*Channel, error) {
 	var channels []*Channel
 	err := DB.Where("id in (?)", ids).Find(&channels).Error

model/log.go

@@ -1,6 +1,7 @@
 package model
 
 import (
+	"context"
 	"fmt"
 	"one-api/common"
 	"os"
@@ -40,6 +41,7 @@ const (
 	LogTypeConsume
 	LogTypeManage
 	LogTypeSystem
+	LogTypeError
 )
 
 func formatUserLogs(logs []*Log) {
@@ -88,6 +90,35 @@ func RecordLog(userId int, logType int, content string) {
 	}
 }
 
+func RecordErrorLog(c *gin.Context, userId int, channelId int, modelName string, tokenName string, content string, tokenId int, useTimeSeconds int,
+	isStream bool, group string, other map[string]interface{}) {
+	common.LogInfo(c, fmt.Sprintf("record error log: userId=%d, channelId=%d, modelName=%s, tokenName=%s, content=%s", userId, channelId, modelName, tokenName, content))
+	username := c.GetString("username")
+	otherStr := common.MapToJsonStr(other)
+	log := &Log{
+		UserId:           userId,
+		Username:         username,
+		CreatedAt:        common.GetTimestamp(),
+		Type:             LogTypeError,
+		Content:          content,
+		PromptTokens:     0,
+		CompletionTokens: 0,
+		TokenName:        tokenName,
+		ModelName:        modelName,
+		Quota:            0,
+		ChannelId:        channelId,
+		TokenId:          tokenId,
+		UseTime:          useTimeSeconds,
+		IsStream:         isStream,
+		Group:            group,
+		Other:            otherStr,
+	}
+	err := LOG_DB.Create(log).Error
+	if err != nil {
+		common.LogError(c, "failed to record log: "+err.Error())
+	}
+}
+
 func RecordConsumeLog(c *gin.Context, userId int, channelId int, promptTokens int, completionTokens int,
 	modelName string, tokenName string, quota int, content string, tokenId int, userQuota int, useTimeSeconds int,
 	isStream bool, group string, other map[string]interface{}) {
@@ -310,7 +341,25 @@ func SumUsedToken(logType int, startTimestamp int64, endTimestamp int64, modelNa
 	return token
 }
 
-func DeleteOldLog(targetTimestamp int64) (int64, error) {
-	result := LOG_DB.Where("created_at < ?", targetTimestamp).Delete(&Log{})
-	return result.RowsAffected, result.Error
+func DeleteOldLog(ctx context.Context, targetTimestamp int64, limit int) (int64, error) {
+	var total int64 = 0
+
+	for {
+		if nil != ctx.Err() {
+			return total, ctx.Err()
+		}
+
+		result := LOG_DB.Where("created_at < ?", targetTimestamp).Limit(limit).Delete(&Log{})
+		if nil != result.Error {
+			return total, result.Error
+		}
+
+		total += result.RowsAffected
+
+		if result.RowsAffected < int64(limit) {
+			break
+		}
+	}
+
+	return total, nil
 }

model/main.go

@@ -1,16 +1,18 @@
 package model
 
 import (
-	"github.com/glebarez/sqlite"
-	"gorm.io/driver/mysql"
-	"gorm.io/driver/postgres"
-	"gorm.io/gorm"
 	"log"
 	"one-api/common"
+	"one-api/constant"
 	"os"
 	"strings"
 	"sync"
 	"time"
+
+	"github.com/glebarez/sqlite"
+	"gorm.io/driver/mysql"
+	"gorm.io/driver/postgres"
+	"gorm.io/gorm"
 )
 
 var groupCol string
@@ -54,13 +56,40 @@ func createRootAccountIfNeed() error {
 	return nil
 }
 
+func CheckSetup() {
+	setup := GetSetup()
+	if setup == nil {
+		// No setup record exists, check if we have a root user
+		if RootUserExists() {
+			common.SysLog("system is not initialized, but root user exists")
+			// Create setup record
+			newSetup := Setup{
+				Version:       common.Version,
+				InitializedAt: time.Now().Unix(),
+			}
+			err := DB.Create(&newSetup).Error
+			if err != nil {
+				common.SysLog("failed to create setup record: " + err.Error())
+			}
+			constant.Setup = true
+		} else {
+			common.SysLog("system is not initialized and no root user exists")
+			constant.Setup = false
+		}
+	} else {
+		// Setup record exists, system is initialized
+		common.SysLog("system is already initialized at: " + time.Unix(setup.InitializedAt, 0).String())
+		constant.Setup = true
+	}
+}
+
 func chooseDB(envName string) (*gorm.DB, error) {
 	defer func() {
 		initCol()
 	}()
 	dsn := os.Getenv(envName)
 	if dsn != "" {
-		if strings.HasPrefix(dsn, "postgres://") {
+		if strings.HasPrefix(dsn, "postgres://") || strings.HasPrefix(dsn, "postgresql://") {
 			// Use PostgreSQL
 			common.SysLog("using PostgreSQL as database")
 			common.UsingPostgreSQL = true
@@ -213,8 +242,9 @@ func migrateDB() error {
 	if err != nil {
 		return err
 	}
+	err = DB.AutoMigrate(&Setup{})
 	common.SysLog("database migrated")
-	err = createRootAccountIfNeed()
+	//err = createRootAccountIfNeed()
 	return err
 }
 

model/option.go

@@ -67,6 +67,7 @@ func InitOptionMap() {
 	common.OptionMap["ServerAddress"] = ""
 	common.OptionMap["WorkerUrl"] = setting.WorkerUrl
 	common.OptionMap["WorkerValidKey"] = setting.WorkerValidKey
+	common.OptionMap["WorkerAllowHttpImageRequestEnabled"] = strconv.FormatBool(setting.WorkerAllowHttpImageRequestEnabled)
 	common.OptionMap["PayAddress"] = ""
 	common.OptionMap["CustomCallbackAddress"] = ""
 	common.OptionMap["EpayId"] = ""
@@ -92,6 +93,7 @@ func InitOptionMap() {
 	common.OptionMap["ModelRequestRateLimitCount"] = strconv.Itoa(setting.ModelRequestRateLimitCount)
 	common.OptionMap["ModelRequestRateLimitDurationMinutes"] = strconv.Itoa(setting.ModelRequestRateLimitDurationMinutes)
 	common.OptionMap["ModelRequestRateLimitSuccessCount"] = strconv.Itoa(setting.ModelRequestRateLimitSuccessCount)
+	common.OptionMap["ModelRequestRateLimitGroup"] = setting.ModelRequestRateLimitGroup2JSONString()
 	common.OptionMap["ModelRatio"] = operation_setting.ModelRatio2JSONString()
 	common.OptionMap["ModelPrice"] = operation_setting.ModelPrice2JSONString()
 	common.OptionMap["CacheRatio"] = operation_setting.CacheRatio2JSONString()
@@ -120,6 +122,9 @@ func InitOptionMap() {
 	common.OptionMap["SensitiveWords"] = setting.SensitiveWordsToString()
 	common.OptionMap["StreamCacheQueueLength"] = strconv.Itoa(setting.StreamCacheQueueLength)
 	common.OptionMap["AutomaticDisableKeywords"] = operation_setting.AutomaticDisableKeywordsToString()
+	common.OptionMap["ApiInfo"] = ""
+	common.OptionMap["UptimeKumaUrl"] = ""
+	common.OptionMap["UptimeKumaSlug"] = ""
 
 	// 自动添加所有注册的模型配置
 	modelConfigs := config.GlobalConfig.ExportAllConfigs()
@@ -256,6 +261,8 @@ func updateOptionMap(key string, value string) (err error) {
 			setting.StopOnSensitiveEnabled = boolValue
 		case "SMTPSSLEnabled":
 			common.SMTPSSLEnabled = boolValue
+		case "WorkerAllowHttpImageRequestEnabled":
+			setting.WorkerAllowHttpImageRequestEnabled = boolValue
 		}
 	}
 	switch key {
@@ -338,6 +345,8 @@ func updateOptionMap(key string, value string) (err error) {
 		setting.ModelRequestRateLimitDurationMinutes, _ = strconv.Atoi(value)
 	case "ModelRequestRateLimitSuccessCount":
 		setting.ModelRequestRateLimitSuccessCount, _ = strconv.Atoi(value)
+	case "ModelRequestRateLimitGroup":
+		err = setting.UpdateModelRequestRateLimitGroupByJSONString(value)
 	case "RetryTimes":
 		common.RetryTimes, _ = strconv.Atoi(value)
 	case "DataExportInterval":

model/setup.go

@@ -0,0 +1,16 @@
+package model
+
+type Setup struct {
+	ID            uint   `json:"id" gorm:"primaryKey"`
+	Version       string `json:"version" gorm:"type:varchar(50);not null"`
+	InitializedAt int64  `json:"initialized_at" gorm:"type:bigint;not null"`
+}
+
+func GetSetup() *Setup {
+	var setup Setup
+	err := DB.First(&setup).Error
+	if err != nil {
+		return nil
+	}
+	return &setup
+}

model/token_cache.go

@@ -19,7 +19,7 @@ func cacheSetToken(token Token) error {
 
 func cacheDeleteToken(key string) error {
 	key = common.GenerateHMAC(key)
-	err := common.RedisHDelObj(fmt.Sprintf("token:%s", key))
+	err := common.RedisDelKey(fmt.Sprintf("token:%s", key))
 	if err != nil {
 		return err
 	}

model/user.go

@@ -9,7 +9,6 @@ import (
 	"strings"
 
 	"github.com/bytedance/gopkg/util/gopool"
-
 	"gorm.io/gorm"
 )
 
@@ -19,11 +18,13 @@ type User struct {
 	Id               int            `json:"id"`
 	Username         string         `json:"username" gorm:"unique;index" validate:"max=12"`
 	Password         string         `json:"password" gorm:"not null;" validate:"min=8,max=20"`
+	OriginalPassword string         `json:"original_password" gorm:"-:all"` // this field is only for Password change verification, don't save it to database!
 	DisplayName      string         `json:"display_name" gorm:"index" validate:"max=20"`
 	Role             int            `json:"role" gorm:"type:int;default:1"`   // admin, common
 	Status           int            `json:"status" gorm:"type:int;default:1"` // enabled, disabled
 	Email            string         `json:"email" gorm:"index" validate:"max=50"`
 	GitHubId         string         `json:"github_id" gorm:"column:github_id;index"`
+	OidcId           string         `json:"oidc_id" gorm:"column:oidc_id;index"`
 	WeChatId         string         `json:"wechat_id" gorm:"column:wechat_id;index"`
 	TelegramId       string         `json:"telegram_id" gorm:"column:telegram_id;index"`
 	VerificationCode string         `json:"verification_code" gorm:"-:all"`                                    // this field is only for Email verification, don't save it to database!
@@ -108,7 +109,7 @@ func CheckUserExistOrDeleted(username string, email string) (bool, error) {
 
 func GetMaxUserId() int {
 	var user User
-	DB.Last(&user)
+	DB.Unscoped().Last(&user)
 	return user.Id
 }
 
@@ -442,6 +443,14 @@ func (user *User) FillUserByGitHubId() error {
 	return nil
 }
 
+func (user *User) FillUserByOidcId() error {
+	if user.OidcId == "" {
+		return errors.New("oidc id 为空！")
+	}
+	DB.Where(User{OidcId: user.OidcId}).First(user)
+	return nil
+}
+
 func (user *User) FillUserByWeChatId() error {
 	if user.WeChatId == "" {
 		return errors.New("WeChat id 为空！")
@@ -473,6 +482,10 @@ func IsGitHubIdAlreadyTaken(githubId string) bool {
 	return DB.Unscoped().Where("github_id = ?", githubId).Find(&User{}).RowsAffected == 1
 }
 
+func IsOidcIdAlreadyTaken(oidcId string) bool {
+	return DB.Where("oidc_id = ?", oidcId).Find(&User{}).RowsAffected == 1
+}
+
 func IsTelegramIdAlreadyTaken(telegramId string) bool {
 	return DB.Unscoped().Where("telegram_id = ?", telegramId).Find(&User{}).RowsAffected == 1
 }
@@ -796,3 +809,12 @@ func (user *User) FillUserByLinuxDOId() error {
 	err := DB.Where("linux_do_id = ?", user.LinuxDOId).First(user).Error
 	return err
 }
+
+func RootUserExists() bool {
+	var user User
+	err := DB.Where("role = ?", common.RoleRootUser).First(&user).Error
+	if err != nil {
+		return false
+	}
+	return true
+}

model/user_cache.go

@@ -3,11 +3,12 @@ package model
 import (
 	"encoding/json"
 	"fmt"
-	"github.com/gin-gonic/gin"
 	"one-api/common"
 	"one-api/constant"
 	"time"
 
+	"github.com/gin-gonic/gin"
+
 	"github.com/bytedance/gopkg/util/gopool"
 )
 
@@ -57,7 +58,7 @@ func invalidateUserCache(userId int) error {
 	if !common.RedisEnabled {
 		return nil
 	}
-	return common.RedisHDelObj(getUserCacheKey(userId))
+	return common.RedisDelKey(getUserCacheKey(userId))
 }
 
 // updateUserCache updates all user cache fields using hash

relay/channel/adapter.go

@@ -1,11 +1,12 @@
 package channel
 
 import (
-	"github.com/gin-gonic/gin"
 	"io"
 	"net/http"
 	"one-api/dto"
 	relaycommon "one-api/relay/common"
+
+	"github.com/gin-gonic/gin"
 )
 
 type Adaptor interface {
@@ -13,15 +14,17 @@ type Adaptor interface {
 	Init(info *relaycommon.RelayInfo)
 	GetRequestURL(info *relaycommon.RelayInfo) (string, error)
 	SetupRequestHeader(c *gin.Context, req *http.Header, info *relaycommon.RelayInfo) error
-	ConvertRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error)
+	ConvertOpenAIRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error)
 	ConvertRerankRequest(c *gin.Context, relayMode int, request dto.RerankRequest) (any, error)
 	ConvertEmbeddingRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.EmbeddingRequest) (any, error)
 	ConvertAudioRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.AudioRequest) (io.Reader, error)
 	ConvertImageRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.ImageRequest) (any, error)
+	ConvertOpenAIResponsesRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.OpenAIResponsesRequest) (any, error)
 	DoRequest(c *gin.Context, info *relaycommon.RelayInfo, requestBody io.Reader) (any, error)
 	DoResponse(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (usage any, err *dto.OpenAIErrorWithStatusCode)
 	GetModelList() []string
 	GetChannelName() string
+	ConvertClaudeRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.ClaudeRequest) (any, error)
 }
 
 type TaskAdaptor interface {

relay/channel/ali/adaptor.go

@@ -3,7 +3,6 @@ package ali
 import (
 	"errors"
 	"fmt"
-	"github.com/gin-gonic/gin"
 	"io"
 	"net/http"
 	"one-api/dto"
@@ -11,11 +10,19 @@ import (
 	"one-api/relay/channel/openai"
 	relaycommon "one-api/relay/common"
 	"one-api/relay/constant"
+
+	"github.com/gin-gonic/gin"
 )
 
 type Adaptor struct {
 }
 
+func (a *Adaptor) ConvertClaudeRequest(*gin.Context, *relaycommon.RelayInfo, *dto.ClaudeRequest) (any, error) {
+	//TODO implement me
+	panic("implement me")
+	return nil, nil
+}
+
 func (a *Adaptor) Init(info *relaycommon.RelayInfo) {
 }
 
@@ -24,8 +31,12 @@ func (a *Adaptor) GetRequestURL(info *relaycommon.RelayInfo) (string, error) {
 	switch info.RelayMode {
 	case constant.RelayModeEmbeddings:
 		fullRequestURL = fmt.Sprintf("%s/api/v1/services/embeddings/text-embedding/text-embedding", info.BaseUrl)
+	case constant.RelayModeRerank:
+		fullRequestURL = fmt.Sprintf("%s/api/v1/services/rerank/text-rerank/text-rerank", info.BaseUrl)
 	case constant.RelayModeImagesGenerations:
 		fullRequestURL = fmt.Sprintf("%s/api/v1/services/aigc/text2image/image-synthesis", info.BaseUrl)
+	case constant.RelayModeCompletions:
+		fullRequestURL = fmt.Sprintf("%s/compatible-mode/v1/completions", info.BaseUrl)
 	default:
 		fullRequestURL = fmt.Sprintf("%s/compatible-mode/v1/chat/completions", info.BaseUrl)
 	}
@@ -44,10 +55,16 @@ func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *rel
 	return nil
 }
 
-func (a *Adaptor) ConvertRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
+func (a *Adaptor) ConvertOpenAIRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
 	if request == nil {
 		return nil, errors.New("request is nil")
 	}
+
+	// fix: ali parameter.enable_thinking must be set to false for non-streaming calls
+	if !info.IsStream {
+		request.EnableThinking = false
+	}
+
 	switch info.RelayMode {
 	default:
 		aliReq := requestOpenAI2Ali(*request)
@@ -61,7 +78,7 @@ func (a *Adaptor) ConvertImageRequest(c *gin.Context, info *relaycommon.RelayInf
 }
 
 func (a *Adaptor) ConvertRerankRequest(c *gin.Context, relayMode int, request dto.RerankRequest) (any, error) {
-	return nil, errors.New("not implemented")
+	return ConvertRerankRequest(request), nil
 }
 
 func (a *Adaptor) ConvertEmbeddingRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.EmbeddingRequest) (any, error) {
@@ -73,6 +90,11 @@ func (a *Adaptor) ConvertAudioRequest(c *gin.Context, info *relaycommon.RelayInf
 	return nil, errors.New("not implemented")
 }
 
+func (a *Adaptor) ConvertOpenAIResponsesRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.OpenAIResponsesRequest) (any, error) {
+	// TODO implement me
+	return nil, errors.New("not implemented")
+}
+
 func (a *Adaptor) DoRequest(c *gin.Context, info *relaycommon.RelayInfo, requestBody io.Reader) (any, error) {
 	return channel.DoApiRequest(a, c, info, requestBody)
 }
@@ -83,11 +105,13 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, info *relaycom
 		err, usage = aliImageHandler(c, resp, info)
 	case constant.RelayModeEmbeddings:
 		err, usage = aliEmbeddingHandler(c, resp)
+	case constant.RelayModeRerank:
+		err, usage = RerankHandler(c, resp, info)
 	default:
 		if info.IsStream {
 			err, usage = openai.OaiStreamHandler(c, resp, info)
 		} else {
-			err, usage = openai.OpenaiHandler(c, resp, info.PromptTokens, info.UpstreamModelName)
+			err, usage = openai.OpenaiHandler(c, resp, info)
 		}
 	}
 	return

relay/channel/ali/constants.go

@@ -1,8 +1,14 @@
 package ali
 
 var ModelList = []string{
-	"qwen-turbo", "qwen-plus", "qwen-max", "qwen-max-longcontext",
+	"qwen-turbo",
+	"qwen-plus",
+	"qwen-max",
+	"qwen-max-longcontext",
+	"qwq-32b",
+	"qwen3-235b-a22b",
 	"text-embedding-v1",
+	"gte-rerank-v2",
 }
 
 var ChannelName = "ali"

relay/channel/ali/dto.go

@@ -1,5 +1,7 @@
 package ali
 
+import "one-api/dto"
+
 type AliMessage struct {
 	Content string `json:"content"`
 	Role    string `json:"role"`
@@ -97,3 +99,28 @@ type AliImageRequest struct {
 	} `json:"parameters,omitempty"`
 	ResponseFormat string `json:"response_format,omitempty"`
 }
+
+type AliRerankParameters struct {
+	TopN            *int  `json:"top_n,omitempty"`
+	ReturnDocuments *bool `json:"return_documents,omitempty"`
+}
+
+type AliRerankInput struct {
+	Query     string `json:"query"`
+	Documents []any  `json:"documents"`
+}
+
+type AliRerankRequest struct {
+	Model      string              `json:"model"`
+	Input      AliRerankInput      `json:"input"`
+	Parameters AliRerankParameters `json:"parameters,omitempty"`
+}
+
+type AliRerankResponse struct {
+	Output struct {
+		Results []dto.RerankResponseResult `json:"results"`
+	} `json:"output"`
+	Usage     AliUsage `json:"usage"`
+	RequestId string   `json:"request_id"`
+	AliError
+}

relay/channel/ali/image.go

@@ -26,8 +26,8 @@ func oaiImage2Ali(request dto.ImageRequest) *AliImageRequest {
 	return &imageRequest
 }
 
-func updateTask(info *relaycommon.RelayInfo, taskID string, key string) (*AliResponse, error, []byte) {
-	url := fmt.Sprintf("/api/v1/tasks/%s", taskID)
+func updateTask(info *relaycommon.RelayInfo, taskID string) (*AliResponse, error, []byte) {
+	url := fmt.Sprintf("%s/api/v1/tasks/%s", info.BaseUrl, taskID)
 
 	var aliResponse AliResponse
 
@@ -36,7 +36,7 @@ func updateTask(info *relaycommon.RelayInfo, taskID string, key string) (*AliRes
 		return &aliResponse, err, nil
 	}
 
-	req.Header.Set("Authorization", "Bearer "+key)
+	req.Header.Set("Authorization", "Bearer "+info.ApiKey)
 
 	client := &http.Client{}
 	resp, err := client.Do(req)
@@ -58,7 +58,7 @@ func updateTask(info *relaycommon.RelayInfo, taskID string, key string) (*AliRes
 	return &response, nil, responseBody
 }
 
-func asyncTaskWait(info *relaycommon.RelayInfo, taskID string, key string) (*AliResponse, []byte, error) {
+func asyncTaskWait(info *relaycommon.RelayInfo, taskID string) (*AliResponse, []byte, error) {
 	waitSeconds := 3
 	step := 0
 	maxStep := 20
@@ -68,7 +68,7 @@ func asyncTaskWait(info *relaycommon.RelayInfo, taskID string, key string) (*Ali
 
 	for {
 		step++
-		rsp, err, body := updateTask(info, taskID, key)
+		rsp, err, body := updateTask(info, taskID)
 		responseBody = body
 		if err != nil {
 			return &taskResponse, responseBody, err
@@ -125,8 +125,6 @@ func responseAli2OpenAIImage(c *gin.Context, response *AliResponse, info *relayc
 }
 
 func aliImageHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (*dto.OpenAIErrorWithStatusCode, *dto.Usage) {
-	apiKey := c.Request.Header.Get("Authorization")
-	apiKey = strings.TrimPrefix(apiKey, "Bearer ")
 	responseFormat := c.GetString("response_format")
 
 	var aliTaskResponse AliResponse
@@ -148,7 +146,7 @@ func aliImageHandler(c *gin.Context, resp *http.Response, info *relaycommon.Rela
 		return service.OpenAIErrorWrapper(errors.New(aliTaskResponse.Message), "ali_async_task_failed", http.StatusInternalServerError), nil
 	}
 
-	aliResponse, _, err := asyncTaskWait(info, aliTaskResponse.Output.TaskId, apiKey)
+	aliResponse, _, err := asyncTaskWait(info, aliTaskResponse.Output.TaskId)
 	if err != nil {
 		return service.OpenAIErrorWrapper(err, "ali_async_task_wait_failed", http.StatusInternalServerError), nil
 	}

relay/channel/ali/rerank.go

@@ -0,0 +1,83 @@
+package ali
+
+import (
+	"encoding/json"
+	"io"
+	"net/http"
+	"one-api/dto"
+	relaycommon "one-api/relay/common"
+	"one-api/service"
+
+	"github.com/gin-gonic/gin"
+)
+
+func ConvertRerankRequest(request dto.RerankRequest) *AliRerankRequest {
+	returnDocuments := request.ReturnDocuments
+	if returnDocuments == nil {
+		t := true
+		returnDocuments = &t
+	}
+	return &AliRerankRequest{
+		Model: request.Model,
+		Input: AliRerankInput{
+			Query:     request.Query,
+			Documents: request.Documents,
+		},
+		Parameters: AliRerankParameters{
+			TopN:            &request.TopN,
+			ReturnDocuments: returnDocuments,
+		},
+	}
+}
+
+func RerankHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (*dto.OpenAIErrorWithStatusCode, *dto.Usage) {
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return service.OpenAIErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = resp.Body.Close()
+	if err != nil {
+		return service.OpenAIErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+
+	var aliResponse AliRerankResponse
+	err = json.Unmarshal(responseBody, &aliResponse)
+	if err != nil {
+		return service.OpenAIErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+
+	if aliResponse.Code != "" {
+		return &dto.OpenAIErrorWithStatusCode{
+			Error: dto.OpenAIError{
+				Message: aliResponse.Message,
+				Type:    aliResponse.Code,
+				Param:   aliResponse.RequestId,
+				Code:    aliResponse.Code,
+			},
+			StatusCode: resp.StatusCode,
+		}, nil
+	}
+
+	usage := dto.Usage{
+		PromptTokens:     aliResponse.Usage.TotalTokens,
+		CompletionTokens: 0,
+		TotalTokens:      aliResponse.Usage.TotalTokens,
+	}
+	rerankResponse := dto.RerankResponse{
+		Results: aliResponse.Output.Results,
+		Usage:   usage,
+	}
+
+	jsonResponse, err := json.Marshal(rerankResponse)
+	if err != nil {
+		return service.OpenAIErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	c.Writer.Header().Set("Content-Type", "application/json")
+	c.Writer.WriteHeader(resp.StatusCode)
+	_, err = c.Writer.Write(jsonResponse)
+	if err != nil {
+		return service.OpenAIErrorWrapper(err, "write_response_body_failed", http.StatusInternalServerError), nil
+	}
+
+	return nil, &usage
+}

relay/channel/ali/text.go

@@ -3,7 +3,6 @@ package ali
 import (
 	"bufio"
 	"encoding/json"
-	"github.com/gin-gonic/gin"
 	"io"
 	"net/http"
 	"one-api/common"
@@ -11,6 +10,8 @@ import (
 	"one-api/relay/helper"
 	"one-api/service"
 	"strings"
+
+	"github.com/gin-gonic/gin"
 )
 
 // https://help.aliyun.com/document_detail/613695.html?spm=a2c4g.2399480.0.0.1adb778fAdzP9w#341800c0f8w0r
@@ -27,9 +28,6 @@ func requestOpenAI2Ali(request dto.GeneralOpenAIRequest) *dto.GeneralOpenAIReque
 }
 
 func embeddingRequestOpenAI2Ali(request dto.EmbeddingRequest) *AliEmbeddingRequest {
-	if request.Model == "" {
-		request.Model = "text-embedding-v1"
-	}
 	return &AliEmbeddingRequest{
 		Model: request.Model,
 		Input: struct {
@@ -64,7 +62,11 @@ func aliEmbeddingHandler(c *gin.Context, resp *http.Response) (*dto.OpenAIErrorW
 		}, nil
 	}
 
-	fullTextResponse := embeddingResponseAli2OpenAI(&aliResponse)
+	model := c.GetString("model")
+	if model == "" {
+		model = "text-embedding-v4"
+	}
+	fullTextResponse := embeddingResponseAli2OpenAI(&aliResponse, model)
 	jsonResponse, err := json.Marshal(fullTextResponse)
 	if err != nil {
 		return service.OpenAIErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
@@ -75,11 +77,11 @@ func aliEmbeddingHandler(c *gin.Context, resp *http.Response) (*dto.OpenAIErrorW
 	return nil, &fullTextResponse.Usage
 }
 
-func embeddingResponseAli2OpenAI(response *AliEmbeddingResponse) *dto.OpenAIEmbeddingResponse {
+func embeddingResponseAli2OpenAI(response *AliEmbeddingResponse, model string) *dto.OpenAIEmbeddingResponse {
 	openAIEmbeddingResponse := dto.OpenAIEmbeddingResponse{
 		Object: "list",
 		Data:   make([]dto.OpenAIEmbeddingResponseItem, 0, len(response.Output.Embeddings)),
-		Model:  "text-embedding-v1",
+		Model:  model,
 		Usage:  dto.Usage{TotalTokens: response.Usage.TotalTokens},
 	}
 
@@ -94,12 +96,11 @@ func embeddingResponseAli2OpenAI(response *AliEmbeddingResponse) *dto.OpenAIEmbe
 }
 
 func responseAli2OpenAI(response *AliResponse) *dto.OpenAITextResponse {
-	content, _ := json.Marshal(response.Output.Text)
 	choice := dto.OpenAITextResponseChoice{
 		Index: 0,
 		Message: dto.Message{
 			Role:    "assistant",
-			Content: content,
+			Content: response.Output.Text,
 		},
 		FinishReason: response.Output.FinishReason,
 	}

relay/channel/api_request.go

@@ -1,15 +1,23 @@
 package channel
 
 import (
+	"context"
 	"errors"
 	"fmt"
-	"github.com/gin-gonic/gin"
-	"github.com/gorilla/websocket"
 	"io"
 	"net/http"
+	common2 "one-api/common"
 	"one-api/relay/common"
 	"one-api/relay/constant"
+	"one-api/relay/helper"
 	"one-api/service"
+	"one-api/setting/operation_setting"
+	"sync"
+	"time"
+
+	"github.com/bytedance/gopkg/util/gopool"
+	"github.com/gin-gonic/gin"
+	"github.com/gorilla/websocket"
 )
 
 func SetupApiRequestHeader(info *common.RelayInfo, c *gin.Context, req *http.Header) {
@@ -31,6 +39,9 @@ func DoApiRequest(a Adaptor, c *gin.Context, info *common.RelayInfo, requestBody
 	if err != nil {
 		return nil, fmt.Errorf("get request url failed: %w", err)
 	}
+	if common2.DebugEnabled {
+		println("fullRequestURL:", fullRequestURL)
+	}
 	req, err := http.NewRequest(c.Request.Method, fullRequestURL, requestBody)
 	if err != nil {
 		return nil, fmt.Errorf("new request failed: %w", err)
@@ -51,6 +62,9 @@ func DoFormRequest(a Adaptor, c *gin.Context, info *common.RelayInfo, requestBod
 	if err != nil {
 		return nil, fmt.Errorf("get request url failed: %w", err)
 	}
+	if common2.DebugEnabled {
+		println("fullRequestURL:", fullRequestURL)
+	}
 	req, err := http.NewRequest(c.Request.Method, fullRequestURL, requestBody)
 	if err != nil {
 		return nil, fmt.Errorf("new request failed: %w", err)
@@ -90,6 +104,105 @@ func DoWssRequest(a Adaptor, c *gin.Context, info *common.RelayInfo, requestBody
 	return targetConn, nil
 }
 
+func startPingKeepAlive(c *gin.Context, pingInterval time.Duration) context.CancelFunc {
+	pingerCtx, stopPinger := context.WithCancel(context.Background())
+
+	gopool.Go(func() {
+		defer func() {
+			// 增加panic恢复处理
+			if r := recover(); r != nil {
+				if common2.DebugEnabled {
+					println("SSE ping goroutine panic recovered:", fmt.Sprintf("%v", r))
+				}
+			}
+			if common2.DebugEnabled {
+				println("SSE ping goroutine stopped.")
+			}
+		}()
+
+		if pingInterval <= 0 {
+			pingInterval = helper.DefaultPingInterval
+		}
+
+		ticker := time.NewTicker(pingInterval)
+		// 确保在任何情况下都清理ticker
+		defer func() {
+			ticker.Stop()
+			if common2.DebugEnabled {
+				println("SSE ping ticker stopped")
+			}
+		}()
+
+		var pingMutex sync.Mutex
+		if common2.DebugEnabled {
+			println("SSE ping goroutine started")
+		}
+
+		// 增加超时控制，防止goroutine长时间运行
+		maxPingDuration := 120 * time.Minute // 最大ping持续时间
+		pingTimeout := time.NewTimer(maxPingDuration)
+		defer pingTimeout.Stop()
+
+		for {
+			select {
+			// 发送 ping 数据
+			case <-ticker.C:
+				if err := sendPingData(c, &pingMutex); err != nil {
+					if common2.DebugEnabled {
+						println("SSE ping error, stopping goroutine:", err.Error())
+					}
+					return
+				}
+			// 收到退出信号
+			case <-pingerCtx.Done():
+				return
+			// request 结束
+			case <-c.Request.Context().Done():
+				return
+			// 超时保护，防止goroutine无限运行
+			case <-pingTimeout.C:
+				if common2.DebugEnabled {
+					println("SSE ping goroutine timeout, stopping")
+				}
+				return
+			}
+		}
+	})
+
+	return stopPinger
+}
+
+func sendPingData(c *gin.Context, mutex *sync.Mutex) error {
+	// 增加超时控制，防止锁死等待
+	done := make(chan error, 1)
+	go func() {
+		mutex.Lock()
+		defer mutex.Unlock()
+
+		err := helper.PingData(c)
+		if err != nil {
+			common2.LogError(c, "SSE ping error: "+err.Error())
+			done <- err
+			return
+		}
+
+		if common2.DebugEnabled {
+			println("SSE ping data sent.")
+		}
+		done <- nil
+	}()
+
+	// 设置发送ping数据的超时时间
+	select {
+	case err := <-done:
+		return err
+	case <-time.After(10 * time.Second):
+		return errors.New("SSE ping data send timeout")
+	case <-c.Request.Context().Done():
+		return errors.New("request context cancelled during ping")
+	}
+}
+
 func doRequest(c *gin.Context, req *http.Request, info *common.RelayInfo) (*http.Response, error) {
 	var client *http.Client
 	var err error
@@ -101,13 +214,36 @@ func doRequest(c *gin.Context, req *http.Request, info *common.RelayInfo) (*http
 	} else {
 		client = service.GetHttpClient()
 	}
+
+	var stopPinger context.CancelFunc
+	if info.IsStream {
+		helper.SetEventStreamHeaders(c)
+		// 处理流式请求的 ping 保活
+		generalSettings := operation_setting.GetGeneralSetting()
+		if generalSettings.PingIntervalEnabled {
+			pingInterval := time.Duration(generalSettings.PingIntervalSeconds) * time.Second
+			stopPinger = startPingKeepAlive(c, pingInterval)
+			// 使用defer确保在任何情况下都能停止ping goroutine
+			defer func() {
+				if stopPinger != nil {
+					stopPinger()
+					if common2.DebugEnabled {
+						println("SSE ping goroutine stopped by defer")
+					}
+				}
+			}()
+		}
+	}
+
 	resp, err := client.Do(req)
+
 	if err != nil {
 		return nil, err
 	}
 	if resp == nil {
 		return nil, errors.New("resp is nil")
 	}
+
 	_ = req.Body.Close()
 	_ = c.Request.Body.Close()
 	return resp, nil

relay/channel/aws/adaptor.go

@@ -2,13 +2,14 @@ package aws
 
 import (
 	"errors"
-	"github.com/gin-gonic/gin"
 	"io"
 	"net/http"
 	"one-api/dto"
 	"one-api/relay/channel/claude"
 	relaycommon "one-api/relay/common"
 	"one-api/setting/model_setting"
+
+	"github.com/gin-gonic/gin"
 )
 
 const (
@@ -20,6 +21,12 @@ type Adaptor struct {
 	RequestMode int
 }
 
+func (a *Adaptor) ConvertClaudeRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.ClaudeRequest) (any, error) {
+	c.Set("request_model", request.Model)
+	c.Set("converted_request", request)
+	return request, nil
+}
+
 func (a *Adaptor) ConvertAudioRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.AudioRequest) (io.Reader, error) {
 	//TODO implement me
 	return nil, errors.New("not implemented")
@@ -43,12 +50,12 @@ func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *rel
 	return nil
 }
 
-func (a *Adaptor) ConvertRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
+func (a *Adaptor) ConvertOpenAIRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
 	if request == nil {
 		return nil, errors.New("request is nil")
 	}
 
-	var claudeReq *claude.ClaudeRequest
+	var claudeReq *dto.ClaudeRequest
 	var err error
 	claudeReq, err = claude.RequestOpenAI2ClaudeMessage(*request)
 	if err != nil {
@@ -68,6 +75,11 @@ func (a *Adaptor) ConvertEmbeddingRequest(c *gin.Context, info *relaycommon.Rela
 	return nil, errors.New("not implemented")
 }
 
+func (a *Adaptor) ConvertOpenAIResponsesRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.OpenAIResponsesRequest) (any, error) {
+	// TODO implement me
+	return nil, errors.New("not implemented")
+}
+
 func (a *Adaptor) DoRequest(c *gin.Context, info *relaycommon.RelayInfo, requestBody io.Reader) (any, error) {
 	return nil, nil
 }

relay/channel/aws/constants.go

@@ -11,6 +11,55 @@ var awsModelIDMap = map[string]string{
 	"claude-3-5-sonnet-20241022": "anthropic.claude-3-5-sonnet-20241022-v2:0",
 	"claude-3-5-haiku-20241022":  "anthropic.claude-3-5-haiku-20241022-v1:0",
 	"claude-3-7-sonnet-20250219": "anthropic.claude-3-7-sonnet-20250219-v1:0",
+	"claude-sonnet-4-20250514":   "anthropic.claude-sonnet-4-20250514-v1:0",
+	"claude-opus-4-20250514":     "anthropic.claude-opus-4-20250514-v1:0",
+}
+
+var awsModelCanCrossRegionMap = map[string]map[string]bool{
+	"anthropic.claude-3-sonnet-20240229-v1:0": {
+		"us": true,
+		"eu": true,
+		"ap": true,
+	},
+	"anthropic.claude-3-opus-20240229-v1:0": {
+		"us": true,
+	},
+	"anthropic.claude-3-haiku-20240307-v1:0": {
+		"us": true,
+		"eu": true,
+		"ap": true,
+	},
+	"anthropic.claude-3-5-sonnet-20240620-v1:0": {
+		"us": true,
+		"eu": true,
+		"ap": true,
+	},
+	"anthropic.claude-3-5-sonnet-20241022-v2:0": {
+		"us": true,
+		"ap": true,
+	},
+	"anthropic.claude-3-5-haiku-20241022-v1:0": {
+		"us": true,
+	},
+	"anthropic.claude-3-7-sonnet-20250219-v1:0": {
+		"us": true,
+		"ap": true,
+		"eu": true,
+	},
+	"anthropic.claude-sonnet-4-20250514-v1:0": {
+		"us": true,
+		"ap": true,
+		"eu": true,
+	},
+	"anthropic.claude-opus-4-20250514-v1:0": {
+		"us": true,
+	},
+}
+
+var awsRegionCrossModelPrefixMap = map[string]string{
+	"us": "us",
+	"eu": "eu",
+	"ap": "apac",
 }
 
 var ChannelName = "aws"

relay/channel/aws/dto.go

@@ -1,25 +1,25 @@
 package aws
 
 import (
-	"one-api/relay/channel/claude"
+	"one-api/dto"
 )
 
 type AwsClaudeRequest struct {
 	// AnthropicVersion should be "bedrock-2023-05-31"
-	AnthropicVersion string                 `json:"anthropic_version"`
-	System           string                 `json:"system,omitempty"`
-	Messages         []claude.ClaudeMessage `json:"messages"`
-	MaxTokens        uint                   `json:"max_tokens,omitempty"`
-	Temperature      *float64               `json:"temperature,omitempty"`
-	TopP             float64                `json:"top_p,omitempty"`
-	TopK             int                    `json:"top_k,omitempty"`
-	StopSequences    []string               `json:"stop_sequences,omitempty"`
-	Tools            any                    `json:"tools,omitempty"`
-	ToolChoice       any                    `json:"tool_choice,omitempty"`
-	Thinking         *claude.Thinking       `json:"thinking,omitempty"`
+	AnthropicVersion string              `json:"anthropic_version"`
+	System           any                 `json:"system,omitempty"`
+	Messages         []dto.ClaudeMessage `json:"messages"`
+	MaxTokens        uint                `json:"max_tokens,omitempty"`
+	Temperature      *float64            `json:"temperature,omitempty"`
+	TopP             float64             `json:"top_p,omitempty"`
+	TopK             int                 `json:"top_k,omitempty"`
+	StopSequences    []string            `json:"stop_sequences,omitempty"`
+	Tools            any                 `json:"tools,omitempty"`
+	ToolChoice       any                 `json:"tool_choice,omitempty"`
+	Thinking         *dto.Thinking       `json:"thinking,omitempty"`
 }
 
-func copyRequest(req *claude.ClaudeRequest) *AwsClaudeRequest {
+func copyRequest(req *dto.ClaudeRequest) *AwsClaudeRequest {
 	return &AwsClaudeRequest{
 		AnthropicVersion: "bedrock-2023-05-31",
 		System:           req.System,

relay/channel/aws/relay-aws.go

@@ -1,21 +1,16 @@
 package aws
 
 import (
-	"bytes"
 	"encoding/json"
 	"fmt"
 	"github.com/gin-gonic/gin"
 	"github.com/pkg/errors"
-	"io"
 	"net/http"
 	"one-api/common"
-	relaymodel "one-api/dto"
+	"one-api/dto"
 	"one-api/relay/channel/claude"
 	relaycommon "one-api/relay/common"
-	"one-api/relay/helper"
-	"one-api/service"
 	"strings"
-	"time"
 
 	"github.com/aws/aws-sdk-go-v2/aws"
 	"github.com/aws/aws-sdk-go-v2/credentials"
@@ -39,15 +34,37 @@ func newAwsClient(c *gin.Context, info *relaycommon.RelayInfo) (*bedrockruntime.
 	return client, nil
 }
 
-func wrapErr(err error) *relaymodel.OpenAIErrorWithStatusCode {
-	return &relaymodel.OpenAIErrorWithStatusCode{
+func wrapErr(err error) *dto.OpenAIErrorWithStatusCode {
+	return &dto.OpenAIErrorWithStatusCode{
 		StatusCode: http.StatusInternalServerError,
-		Error: relaymodel.OpenAIError{
+		Error: dto.OpenAIError{
 			Message: fmt.Sprintf("%s", err.Error()),
 		},
 	}
 }
 
+func awsRegionPrefix(awsRegionId string) string {
+	parts := strings.Split(awsRegionId, "-")
+	regionPrefix := ""
+	if len(parts) > 0 {
+		regionPrefix = parts[0]
+	}
+	return regionPrefix
+}
+
+func awsModelCanCrossRegion(awsModelId, awsRegionPrefix string) bool {
+	regionSet, exists := awsModelCanCrossRegionMap[awsModelId]
+	return exists && regionSet[awsRegionPrefix]
+}
+
+func awsModelCrossRegion(awsModelId, awsRegionPrefix string) string {
+	modelPrefix, find := awsRegionCrossModelPrefixMap[awsRegionPrefix]
+	if !find {
+		return awsModelId
+	}
+	return modelPrefix + "." + awsModelId
+}
+
 func awsModelID(requestModel string) (string, error) {
 	if awsModelID, ok := awsModelIDMap[requestModel]; ok {
 		return awsModelID, nil
@@ -56,7 +73,7 @@ func awsModelID(requestModel string) (string, error) {
 	return requestModel, nil
 }
 
-func awsHandler(c *gin.Context, info *relaycommon.RelayInfo, requestMode int) (*relaymodel.OpenAIErrorWithStatusCode, *relaymodel.Usage) {
+func awsHandler(c *gin.Context, info *relaycommon.RelayInfo, requestMode int) (*dto.OpenAIErrorWithStatusCode, *dto.Usage) {
 	awsCli, err := newAwsClient(c, info)
 	if err != nil {
 		return wrapErr(errors.Wrap(err, "newAwsClient")), nil
@@ -67,6 +84,12 @@ func awsHandler(c *gin.Context, info *relaycommon.RelayInfo, requestMode int) (*
 		return wrapErr(errors.Wrap(err, "awsModelID")), nil
 	}
 
+	awsRegionPrefix := awsRegionPrefix(awsCli.Options().Region)
+	canCrossRegion := awsModelCanCrossRegion(awsModelId, awsRegionPrefix)
+	if canCrossRegion {
+		awsModelId = awsModelCrossRegion(awsModelId, awsRegionPrefix)
+	}
+
 	awsReq := &bedrockruntime.InvokeModelInput{
 		ModelId:     aws.String(awsModelId),
 		Accept:      aws.String("application/json"),
@@ -77,7 +100,7 @@ func awsHandler(c *gin.Context, info *relaycommon.RelayInfo, requestMode int) (*
 	if !ok {
 		return wrapErr(errors.New("request not found")), nil
 	}
-	claudeReq := claudeReq_.(*claude.ClaudeRequest)
+	claudeReq := claudeReq_.(*dto.ClaudeRequest)
 	awsClaudeReq := copyRequest(claudeReq)
 	awsReq.Body, err = json.Marshal(awsClaudeReq)
 	if err != nil {
@@ -89,25 +112,19 @@ func awsHandler(c *gin.Context, info *relaycommon.RelayInfo, requestMode int) (*
 		return wrapErr(errors.Wrap(err, "InvokeModel")), nil
 	}
 
-	claudeResponse := new(claude.ClaudeResponse)
-	err = json.Unmarshal(awsResp.Body, claudeResponse)
-	if err != nil {
-		return wrapErr(errors.Wrap(err, "unmarshal response")), nil
+	claudeInfo := &claude.ClaudeResponseInfo{
+		ResponseId:   fmt.Sprintf("chatcmpl-%s", common.GetUUID()),
+		Created:      common.GetTimestamp(),
+		Model:        info.UpstreamModelName,
+		ResponseText: strings.Builder{},
+		Usage:        &dto.Usage{},
 	}
 
-	openaiResp := claude.ResponseClaude2OpenAI(requestMode, claudeResponse)
-	usage := relaymodel.Usage{
-		PromptTokens:     claudeResponse.Usage.InputTokens,
-		CompletionTokens: claudeResponse.Usage.OutputTokens,
-		TotalTokens:      claudeResponse.Usage.InputTokens + claudeResponse.Usage.OutputTokens,
-	}
-	openaiResp.Usage = usage
-
-	c.JSON(http.StatusOK, openaiResp)
-	return nil, &usage
+	claude.HandleClaudeResponseData(c, info, claudeInfo, awsResp.Body, RequestModeMessage)
+	return nil, claudeInfo.Usage
 }
 
-func awsStreamHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo, requestMode int) (*relaymodel.OpenAIErrorWithStatusCode, *relaymodel.Usage) {
+func awsStreamHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo, requestMode int) (*dto.OpenAIErrorWithStatusCode, *dto.Usage) {
 	awsCli, err := newAwsClient(c, info)
 	if err != nil {
 		return wrapErr(errors.Wrap(err, "newAwsClient")), nil
@@ -118,6 +135,12 @@ func awsStreamHandler(c *gin.Context, resp *http.Response, info *relaycommon.Rel
 		return wrapErr(errors.Wrap(err, "awsModelID")), nil
 	}
 
+	awsRegionPrefix := awsRegionPrefix(awsCli.Options().Region)
+	canCrossRegion := awsModelCanCrossRegion(awsModelId, awsRegionPrefix)
+	if canCrossRegion {
+		awsModelId = awsModelCrossRegion(awsModelId, awsRegionPrefix)
+	}
+
 	awsReq := &bedrockruntime.InvokeModelWithResponseStreamInput{
 		ModelId:     aws.String(awsModelId),
 		Accept:      aws.String("application/json"),
@@ -128,7 +151,7 @@ func awsStreamHandler(c *gin.Context, resp *http.Response, info *relaycommon.Rel
 	if !ok {
 		return wrapErr(errors.New("request not found")), nil
 	}
-	claudeReq := claudeReq_.(*claude.ClaudeRequest)
+	claudeReq := claudeReq_.(*dto.ClaudeRequest)
 
 	awsClaudeReq := copyRequest(claudeReq)
 	awsReq.Body, err = json.Marshal(awsClaudeReq)
@@ -143,79 +166,31 @@ func awsStreamHandler(c *gin.Context, resp *http.Response, info *relaycommon.Rel
 	stream := awsResp.GetStream()
 	defer stream.Close()
 
-	c.Writer.Header().Set("Content-Type", "text/event-stream")
-	var usage relaymodel.Usage
-	var id string
-	var model string
-	isFirst := true
-	createdTime := common.GetTimestamp()
-	c.Stream(func(w io.Writer) bool {
-		event, ok := <-stream.Events()
-		if !ok {
-			return false
-		}
+	claudeInfo := &claude.ClaudeResponseInfo{
+		ResponseId:   fmt.Sprintf("chatcmpl-%s", common.GetUUID()),
+		Created:      common.GetTimestamp(),
+		Model:        info.UpstreamModelName,
+		ResponseText: strings.Builder{},
+		Usage:        &dto.Usage{},
+	}
 
+	for event := range stream.Events() {
 		switch v := event.(type) {
 		case *types.ResponseStreamMemberChunk:
-			if isFirst {
-				isFirst = false
-				info.FirstResponseTime = time.Now()
+			info.SetFirstResponseTime()
+			respErr := claude.HandleStreamResponseData(c, info, claudeInfo, string(v.Value.Bytes), RequestModeMessage)
+			if respErr != nil {
+				return respErr, nil
 			}
-			claudeResp := new(claude.ClaudeResponse)
-			err := json.NewDecoder(bytes.NewReader(v.Value.Bytes)).Decode(claudeResp)
-			if err != nil {
-				common.SysError("error unmarshalling stream response: " + err.Error())
-				return false
-			}
-
-			response, claudeUsage := claude.StreamResponseClaude2OpenAI(requestMode, claudeResp)
-			if claudeUsage != nil {
-				usage.PromptTokens += claudeUsage.InputTokens
-				usage.CompletionTokens += claudeUsage.OutputTokens
-			}
-
-			if response == nil {
-				return true
-			}
-
-			if response.Id != "" {
-				id = response.Id
-			}
-			if response.Model != "" {
-				model = response.Model
-			}
-			response.Created = createdTime
-			response.Id = id
-			response.Model = model
-
-			jsonStr, err := json.Marshal(response)
-			if err != nil {
-				common.SysError("error marshalling stream response: " + err.Error())
-				return true
-			}
-			c.Render(-1, common.CustomEvent{Data: "data: " + string(jsonStr)})
-			return true
 		case *types.UnknownUnionMember:
 			fmt.Println("unknown tag:", v.Tag)
-			return false
+			return wrapErr(errors.New("unknown response type")), nil
 		default:
 			fmt.Println("union is nil or unknown type")
-			return false
-		}
-	})
-	if info.ShouldIncludeUsage {
-		response := helper.GenerateFinalUsageResponse(id, createdTime, info.UpstreamModelName, usage)
-		err := helper.ObjectData(c, response)
-		if err != nil {
-			common.SysError("send final response failed: " + err.Error())
+			return wrapErr(errors.New("nil or unknown response type")), nil
 		}
 	}
-	helper.Done(c)
-	if resp != nil {
-		err = resp.Body.Close()
-		if err != nil {
-			return service.OpenAIErrorWrapperLocal(err, "close_response_body_failed", http.StatusInternalServerError), nil
-		}
-	}
-	return nil, &usage
+
+	claude.HandleStreamFinalResponse(c, info, claudeInfo, RequestModeMessage)
+	return nil, claudeInfo.Usage
 }

relay/channel/baidu/adaptor.go

@@ -3,7 +3,6 @@ package baidu
 import (
 	"errors"
 	"fmt"
-	"github.com/gin-gonic/gin"
 	"io"
 	"net/http"
 	"one-api/dto"
@@ -11,11 +10,19 @@ import (
 	relaycommon "one-api/relay/common"
 	"one-api/relay/constant"
 	"strings"
+
+	"github.com/gin-gonic/gin"
 )
 
 type Adaptor struct {
 }
 
+func (a *Adaptor) ConvertClaudeRequest(*gin.Context, *relaycommon.RelayInfo, *dto.ClaudeRequest) (any, error) {
+	//TODO implement me
+	panic("implement me")
+	return nil, nil
+}
+
 func (a *Adaptor) ConvertAudioRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.AudioRequest) (io.Reader, error) {
 	//TODO implement me
 	return nil, errors.New("not implemented")
@@ -104,7 +111,7 @@ func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *rel
 	return nil
 }
 
-func (a *Adaptor) ConvertRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
+func (a *Adaptor) ConvertOpenAIRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
 	if request == nil {
 		return nil, errors.New("request is nil")
 	}
@@ -124,6 +131,11 @@ func (a *Adaptor) ConvertEmbeddingRequest(c *gin.Context, info *relaycommon.Rela
 	return baiduEmbeddingRequest, nil
 }
 
+func (a *Adaptor) ConvertOpenAIResponsesRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.OpenAIResponsesRequest) (any, error) {
+	// TODO implement me
+	return nil, errors.New("not implemented")
+}
+
 func (a *Adaptor) DoRequest(c *gin.Context, info *relaycommon.RelayInfo, requestBody io.Reader) (any, error) {
 	return channel.DoApiRequest(a, c, info, requestBody)
 }

relay/channel/baidu/relay-baidu.go

@@ -53,12 +53,11 @@ func requestOpenAI2Baidu(request dto.GeneralOpenAIRequest) *BaiduChatRequest {
 }
 
 func responseBaidu2OpenAI(response *BaiduChatResponse) *dto.OpenAITextResponse {
-	content, _ := json.Marshal(response.Result)
 	choice := dto.OpenAITextResponseChoice{
 		Index: 0,
 		Message: dto.Message{
 			Role:    "assistant",
-			Content: content,
+			Content: response.Result,
 		},
 		FinishReason: "stop",
 	}

relay/channel/baidu_v2/adaptor.go

@@ -3,18 +3,26 @@ package baidu_v2
 import (
 	"errors"
 	"fmt"
-	"github.com/gin-gonic/gin"
 	"io"
 	"net/http"
 	"one-api/dto"
 	"one-api/relay/channel"
 	"one-api/relay/channel/openai"
 	relaycommon "one-api/relay/common"
+	"strings"
+
+	"github.com/gin-gonic/gin"
 )
 
 type Adaptor struct {
 }
 
+func (a *Adaptor) ConvertClaudeRequest(*gin.Context, *relaycommon.RelayInfo, *dto.ClaudeRequest) (any, error) {
+	//TODO implement me
+	panic("implement me")
+	return nil, nil
+}
+
 func (a *Adaptor) ConvertAudioRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.AudioRequest) (io.Reader, error) {
 	//TODO implement me
 	return nil, errors.New("not implemented")
@@ -38,10 +46,22 @@ func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *rel
 	return nil
 }
 
-func (a *Adaptor) ConvertRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
+func (a *Adaptor) ConvertOpenAIRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
 	if request == nil {
 		return nil, errors.New("request is nil")
 	}
+	if strings.HasSuffix(info.UpstreamModelName, "-search") {
+		info.UpstreamModelName = strings.TrimSuffix(info.UpstreamModelName, "-search")
+		request.Model = info.UpstreamModelName
+		toMap := request.ToMap()
+		toMap["web_search"] = map[string]any{
+			"enable":          true,
+			"enable_citation": true,
+			"enable_trace":    true,
+			"enable_status":   false,
+		}
+		return toMap, nil
+	}
 	return request, nil
 }
 
@@ -54,6 +74,11 @@ func (a *Adaptor) ConvertEmbeddingRequest(c *gin.Context, info *relaycommon.Rela
 	return nil, errors.New("not implemented")
 }
 
+func (a *Adaptor) ConvertOpenAIResponsesRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.OpenAIResponsesRequest) (any, error) {
+	// TODO implement me
+	return nil, errors.New("not implemented")
+}
+
 func (a *Adaptor) DoRequest(c *gin.Context, info *relaycommon.RelayInfo, requestBody io.Reader) (any, error) {
 	return channel.DoApiRequest(a, c, info, requestBody)
 }
@@ -62,7 +87,7 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, info *relaycom
 	if info.IsStream {
 		err, usage = openai.OaiStreamHandler(c, resp, info)
 	} else {
-		err, usage = openai.OpenaiHandler(c, resp, info.PromptTokens, info.UpstreamModelName)
+		err, usage = openai.OpenaiHandler(c, resp, info)
 	}
 	return
 }

relay/channel/claude/adaptor.go

@@ -3,7 +3,6 @@ package claude
 import (
 	"errors"
 	"fmt"
-	"github.com/gin-gonic/gin"
 	"io"
 	"net/http"
 	"one-api/dto"
@@ -11,6 +10,8 @@ import (
 	relaycommon "one-api/relay/common"
 	"one-api/setting/model_setting"
 	"strings"
+
+	"github.com/gin-gonic/gin"
 )
 
 const (
@@ -22,6 +23,10 @@ type Adaptor struct {
 	RequestMode int
 }
 
+func (a *Adaptor) ConvertClaudeRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.ClaudeRequest) (any, error) {
+	return request, nil
+}
+
 func (a *Adaptor) ConvertAudioRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.AudioRequest) (io.Reader, error) {
 	//TODO implement me
 	return nil, errors.New("not implemented")
@@ -33,10 +38,10 @@ func (a *Adaptor) ConvertImageRequest(c *gin.Context, info *relaycommon.RelayInf
 }
 
 func (a *Adaptor) Init(info *relaycommon.RelayInfo) {
-	if strings.HasPrefix(info.UpstreamModelName, "claude-3") {
-		a.RequestMode = RequestModeMessage
-	} else {
+	if strings.HasPrefix(info.UpstreamModelName, "claude-2") || strings.HasPrefix(info.UpstreamModelName, "claude-instant") {
 		a.RequestMode = RequestModeCompletion
+	} else {
+		a.RequestMode = RequestModeMessage
 	}
 }
 
@@ -60,7 +65,7 @@ func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *rel
 	return nil
 }
 
-func (a *Adaptor) ConvertRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
+func (a *Adaptor) ConvertOpenAIRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
 	if request == nil {
 		return nil, errors.New("request is nil")
 	}
@@ -80,6 +85,11 @@ func (a *Adaptor) ConvertEmbeddingRequest(c *gin.Context, info *relaycommon.Rela
 	return nil, errors.New("not implemented")
 }
 
+func (a *Adaptor) ConvertOpenAIResponsesRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.OpenAIResponsesRequest) (any, error) {
+	// TODO implement me
+	return nil, errors.New("not implemented")
+}
+
 func (a *Adaptor) DoRequest(c *gin.Context, info *relaycommon.RelayInfo, requestBody io.Reader) (any, error) {
 	return channel.DoApiRequest(a, c, info, requestBody)
 }

relay/channel/claude/constants.go

@@ -13,6 +13,10 @@ var ModelList = []string{
 	"claude-3-5-sonnet-20241022",
 	"claude-3-7-sonnet-20250219",
 	"claude-3-7-sonnet-20250219-thinking",
+	"claude-sonnet-4-20250514",
+	"claude-sonnet-4-20250514-thinking",
+	"claude-opus-4-20250514",
+	"claude-opus-4-20250514-thinking",
 }
 
 var ChannelName = "claude"

relay/channel/claude/dto.go

@@ -1,94 +1,95 @@
 package claude
 
-type ClaudeMetadata struct {
-	UserId string `json:"user_id"`
-}
-
-type ClaudeMediaMessage struct {
-	Type        string               `json:"type"`
-	Text        string               `json:"text,omitempty"`
-	Source      *ClaudeMessageSource `json:"source,omitempty"`
-	Usage       *ClaudeUsage         `json:"usage,omitempty"`
-	StopReason  *string              `json:"stop_reason,omitempty"`
-	PartialJson string               `json:"partial_json,omitempty"`
-	Thinking    string               `json:"thinking,omitempty"`
-	Signature   string               `json:"signature,omitempty"`
-	Delta       string               `json:"delta,omitempty"`
-	// tool_calls
-	Id        string `json:"id,omitempty"`
-	Name      string `json:"name,omitempty"`
-	Input     any    `json:"input,omitempty"`
-	Content   string `json:"content,omitempty"`
-	ToolUseId string `json:"tool_use_id,omitempty"`
-}
-
-type ClaudeMessageSource struct {
-	Type      string `json:"type"`
-	MediaType string `json:"media_type"`
-	Data      string `json:"data"`
-}
-
-type ClaudeMessage struct {
-	Role    string `json:"role"`
-	Content any    `json:"content"`
-}
-
-type Tool struct {
-	Name        string                 `json:"name"`
-	Description string                 `json:"description,omitempty"`
-	InputSchema map[string]interface{} `json:"input_schema"`
-}
-
-type InputSchema struct {
-	Type       string `json:"type"`
-	Properties any    `json:"properties,omitempty"`
-	Required   any    `json:"required,omitempty"`
-}
-
-type ClaudeRequest struct {
-	Model             string          `json:"model"`
-	Prompt            string          `json:"prompt,omitempty"`
-	System            string          `json:"system,omitempty"`
-	Messages          []ClaudeMessage `json:"messages,omitempty"`
-	MaxTokens         uint            `json:"max_tokens,omitempty"`
-	MaxTokensToSample uint            `json:"max_tokens_to_sample,omitempty"`
-	StopSequences     []string        `json:"stop_sequences,omitempty"`
-	Temperature       *float64        `json:"temperature,omitempty"`
-	TopP              float64         `json:"top_p,omitempty"`
-	TopK              int             `json:"top_k,omitempty"`
-	//ClaudeMetadata    `json:"metadata,omitempty"`
-	Stream     bool      `json:"stream,omitempty"`
-	Tools      any       `json:"tools,omitempty"`
-	ToolChoice any       `json:"tool_choice,omitempty"`
-	Thinking   *Thinking `json:"thinking,omitempty"`
-}
-
-type Thinking struct {
-	Type         string `json:"type"`
-	BudgetTokens int    `json:"budget_tokens"`
-}
-
-type ClaudeError struct {
-	Type    string `json:"type"`
-	Message string `json:"message"`
-}
-
-type ClaudeResponse struct {
-	Id           string               `json:"id"`
-	Type         string               `json:"type"`
-	Content      []ClaudeMediaMessage `json:"content"`
-	Completion   string               `json:"completion"`
-	StopReason   string               `json:"stop_reason"`
-	Model        string               `json:"model"`
-	Error        ClaudeError          `json:"error"`
-	Usage        ClaudeUsage          `json:"usage"`
-	Index        int                  `json:"index"` // stream only
-	ContentBlock *ClaudeMediaMessage  `json:"content_block"`
-	Delta        *ClaudeMediaMessage  `json:"delta"`   // stream only
-	Message      *ClaudeResponse      `json:"message"` // stream only: message_start
-}
-
-type ClaudeUsage struct {
-	InputTokens  int `json:"input_tokens"`
-	OutputTokens int `json:"output_tokens"`
-}
+//
+//type ClaudeMetadata struct {
+//	UserId string `json:"user_id"`
+//}
+//
+//type ClaudeMediaMessage struct {
+//	Type        string               `json:"type"`
+//	Text        string               `json:"text,omitempty"`
+//	Source      *ClaudeMessageSource `json:"source,omitempty"`
+//	Usage       *ClaudeUsage         `json:"usage,omitempty"`
+//	StopReason  *string              `json:"stop_reason,omitempty"`
+//	PartialJson string               `json:"partial_json,omitempty"`
+//	Thinking    string               `json:"thinking,omitempty"`
+//	Signature   string               `json:"signature,omitempty"`
+//	Delta       string               `json:"delta,omitempty"`
+//	// tool_calls
+//	Id        string `json:"id,omitempty"`
+//	Name      string `json:"name,omitempty"`
+//	Input     any    `json:"input,omitempty"`
+//	Content   string `json:"content,omitempty"`
+//	ToolUseId string `json:"tool_use_id,omitempty"`
+//}
+//
+//type ClaudeMessageSource struct {
+//	Type      string `json:"type"`
+//	MediaType string `json:"media_type"`
+//	Data      string `json:"data"`
+//}
+//
+//type ClaudeMessage struct {
+//	Role    string `json:"role"`
+//	Content any    `json:"content"`
+//}
+//
+//type Tool struct {
+//	Name        string                 `json:"name"`
+//	Description string                 `json:"description,omitempty"`
+//	InputSchema map[string]interface{} `json:"input_schema"`
+//}
+//
+//type InputSchema struct {
+//	Type       string `json:"type"`
+//	Properties any    `json:"properties,omitempty"`
+//	Required   any    `json:"required,omitempty"`
+//}
+//
+//type ClaudeRequest struct {
+//	Model             string          `json:"model"`
+//	Prompt            string          `json:"prompt,omitempty"`
+//	System            string          `json:"system,omitempty"`
+//	Messages          []ClaudeMessage `json:"messages,omitempty"`
+//	MaxTokens         uint            `json:"max_tokens,omitempty"`
+//	MaxTokensToSample uint            `json:"max_tokens_to_sample,omitempty"`
+//	StopSequences     []string        `json:"stop_sequences,omitempty"`
+//	Temperature       *float64        `json:"temperature,omitempty"`
+//	TopP              float64         `json:"top_p,omitempty"`
+//	TopK              int             `json:"top_k,omitempty"`
+//	//ClaudeMetadata    `json:"metadata,omitempty"`
+//	Stream     bool      `json:"stream,omitempty"`
+//	Tools      any       `json:"tools,omitempty"`
+//	ToolChoice any       `json:"tool_choice,omitempty"`
+//	Thinking   *Thinking `json:"thinking,omitempty"`
+//}
+//
+//type Thinking struct {
+//	Type         string `json:"type"`
+//	BudgetTokens int    `json:"budget_tokens"`
+//}
+//
+//type ClaudeError struct {
+//	Type    string `json:"type"`
+//	Message string `json:"message"`
+//}
+//
+//type ClaudeResponse struct {
+//	Id           string               `json:"id"`
+//	Type         string               `json:"type"`
+//	Content      []ClaudeMediaMessage `json:"content"`
+//	Completion   string               `json:"completion"`
+//	StopReason   string               `json:"stop_reason"`
+//	Model        string               `json:"model"`
+//	Error        ClaudeError          `json:"error"`
+//	Usage        ClaudeUsage          `json:"usage"`
+//	Index        int                  `json:"index"` // stream only
+//	ContentBlock *ClaudeMediaMessage  `json:"content_block"`
+//	Delta        *ClaudeMediaMessage  `json:"delta"`   // stream only
+//	Message      *ClaudeResponse      `json:"message"` // stream only: message_start
+//}
+//
+//type ClaudeUsage struct {
+//	InputTokens  int `json:"input_tokens"`
+//	OutputTokens int `json:"output_tokens"`
+//}

relay/channel/claude/relay-claude.go

@@ -24,14 +24,16 @@ func stopReasonClaude2OpenAI(reason string) string {
 		return "stop"
 	case "max_tokens":
 		return "max_tokens"
+	case "tool_use":
+		return "tool_calls"
 	default:
 		return reason
 	}
 }
 
-func RequestOpenAI2ClaudeComplete(textRequest dto.GeneralOpenAIRequest) *ClaudeRequest {
+func RequestOpenAI2ClaudeComplete(textRequest dto.GeneralOpenAIRequest) *dto.ClaudeRequest {
 
-	claudeRequest := ClaudeRequest{
+	claudeRequest := dto.ClaudeRequest{
 		Model:         textRequest.Model,
 		Prompt:        "",
 		StopSequences: nil,
@@ -46,9 +48,9 @@ func RequestOpenAI2ClaudeComplete(textRequest dto.GeneralOpenAIRequest) *ClaudeR
 	prompt := ""
 	for _, message := range textRequest.Messages {
 		if message.Role == "user" {
-			prompt += fmt.Sprintf("\n\nHuman: %s", message.Content)
+			prompt += fmt.Sprintf("\n\nHuman: %s", message.StringContent())
 		} else if message.Role == "assistant" {
-			prompt += fmt.Sprintf("\n\nAssistant: %s", message.Content)
+			prompt += fmt.Sprintf("\n\nAssistant: %s", message.StringContent())
 		} else if message.Role == "system" {
 			if prompt == "" {
 				prompt = message.StringContent()
@@ -60,17 +62,19 @@ func RequestOpenAI2ClaudeComplete(textRequest dto.GeneralOpenAIRequest) *ClaudeR
 	return &claudeRequest
 }
 
-func RequestOpenAI2ClaudeMessage(textRequest dto.GeneralOpenAIRequest) (*ClaudeRequest, error) {
-	claudeTools := make([]Tool, 0, len(textRequest.Tools))
+func RequestOpenAI2ClaudeMessage(textRequest dto.GeneralOpenAIRequest) (*dto.ClaudeRequest, error) {
+	claudeTools := make([]dto.Tool, 0, len(textRequest.Tools))
 
 	for _, tool := range textRequest.Tools {
 		if params, ok := tool.Function.Parameters.(map[string]any); ok {
-			claudeTool := Tool{
+			claudeTool := dto.Tool{
 				Name:        tool.Function.Name,
 				Description: tool.Function.Description,
 			}
 			claudeTool.InputSchema = make(map[string]interface{})
-			claudeTool.InputSchema["type"] = params["type"].(string)
+			if params["type"] != nil {
+				claudeTool.InputSchema["type"] = params["type"].(string)
+			}
 			claudeTool.InputSchema["properties"] = params["properties"]
 			claudeTool.InputSchema["required"] = params["required"]
 			for s, a := range params {
@@ -83,7 +87,7 @@ func RequestOpenAI2ClaudeMessage(textRequest dto.GeneralOpenAIRequest) (*ClaudeR
 		}
 	}
 
-	claudeRequest := ClaudeRequest{
+	claudeRequest := dto.ClaudeRequest{
 		Model:         textRequest.Model,
 		MaxTokens:     textRequest.MaxTokens,
 		StopSequences: nil,
@@ -107,7 +111,7 @@ func RequestOpenAI2ClaudeMessage(textRequest dto.GeneralOpenAIRequest) (*ClaudeR
 		}
 
 		// BudgetTokens 为 max_tokens 的 80%
-		claudeRequest.Thinking = &Thinking{
+		claudeRequest.Thinking = &dto.Thinking{
 			Type:         "enabled",
 			BudgetTokens: int(float64(claudeRequest.MaxTokens) * model_setting.GetClaudeSettings().ThinkingAdapterBudgetTokensPercentage),
 		}
@@ -151,21 +155,19 @@ func RequestOpenAI2ClaudeMessage(textRequest dto.GeneralOpenAIRequest) (*ClaudeR
 		}
 		if lastMessage.Role == message.Role && lastMessage.Role != "tool" {
 			if lastMessage.IsStringContent() && message.IsStringContent() {
-				content, _ := json.Marshal(strings.Trim(fmt.Sprintf("%s %s", lastMessage.StringContent(), message.StringContent()), "\""))
-				fmtMessage.Content = content
+				fmtMessage.SetStringContent(strings.Trim(fmt.Sprintf("%s %s", lastMessage.StringContent(), message.StringContent()), "\""))
 				// delete last message
 				formatMessages = formatMessages[:len(formatMessages)-1]
 			}
 		}
 		if fmtMessage.Content == nil {
-			content, _ := json.Marshal("...")
-			fmtMessage.Content = content
+			fmtMessage.SetStringContent("...")
 		}
 		formatMessages = append(formatMessages, fmtMessage)
 		lastMessage = fmtMessage
 	}
 
-	claudeMessages := make([]ClaudeMessage, 0)
+	claudeMessages := make([]dto.ClaudeMessage, 0)
 	isFirstMessage := true
 	for _, message := range formatMessages {
 		if message.Role == "system" {
@@ -186,63 +188,63 @@ func RequestOpenAI2ClaudeMessage(textRequest dto.GeneralOpenAIRequest) (*ClaudeR
 				isFirstMessage = false
 				if message.Role != "user" {
 					// fix: first message is assistant, add user message
-					claudeMessage := ClaudeMessage{
+					claudeMessage := dto.ClaudeMessage{
 						Role: "user",
-						Content: []ClaudeMediaMessage{
+						Content: []dto.ClaudeMediaMessage{
 							{
 								Type: "text",
-								Text: "...",
+								Text: common.GetPointer[string]("..."),
 							},
 						},
 					}
 					claudeMessages = append(claudeMessages, claudeMessage)
 				}
 			}
-			claudeMessage := ClaudeMessage{
+			claudeMessage := dto.ClaudeMessage{
 				Role: message.Role,
 			}
 			if message.Role == "tool" {
 				if len(claudeMessages) > 0 && claudeMessages[len(claudeMessages)-1].Role == "user" {
 					lastMessage := claudeMessages[len(claudeMessages)-1]
 					if content, ok := lastMessage.Content.(string); ok {
-						lastMessage.Content = []ClaudeMediaMessage{
+						lastMessage.Content = []dto.ClaudeMediaMessage{
 							{
 								Type: "text",
-								Text: content,
+								Text: common.GetPointer[string](content),
 							},
 						}
 					}
-					lastMessage.Content = append(lastMessage.Content.([]ClaudeMediaMessage), ClaudeMediaMessage{
+					lastMessage.Content = append(lastMessage.Content.([]dto.ClaudeMediaMessage), dto.ClaudeMediaMessage{
 						Type:      "tool_result",
 						ToolUseId: message.ToolCallId,
-						Content:   message.StringContent(),
+						Content:   message.Content,
 					})
 					claudeMessages[len(claudeMessages)-1] = lastMessage
 					continue
 				} else {
 					claudeMessage.Role = "user"
-					claudeMessage.Content = []ClaudeMediaMessage{
+					claudeMessage.Content = []dto.ClaudeMediaMessage{
 						{
 							Type:      "tool_result",
 							ToolUseId: message.ToolCallId,
-							Content:   message.StringContent(),
+							Content:   message.Content,
 						},
 					}
 				}
 			} else if message.IsStringContent() && message.ToolCalls == nil {
 				claudeMessage.Content = message.StringContent()
 			} else {
-				claudeMediaMessages := make([]ClaudeMediaMessage, 0)
+				claudeMediaMessages := make([]dto.ClaudeMediaMessage, 0)
 				for _, mediaMessage := range message.ParseContent() {
-					claudeMediaMessage := ClaudeMediaMessage{
+					claudeMediaMessage := dto.ClaudeMediaMessage{
 						Type: mediaMessage.Type,
 					}
 					if mediaMessage.Type == "text" {
-						claudeMediaMessage.Text = mediaMessage.Text
+						claudeMediaMessage.Text = common.GetPointer[string](mediaMessage.Text)
 					} else {
-						imageUrl := mediaMessage.ImageUrl.(dto.MessageImageUrl)
+						imageUrl := mediaMessage.GetImageMedia()
 						claudeMediaMessage.Type = "image"
-						claudeMediaMessage.Source = &ClaudeMessageSource{
+						claudeMediaMessage.Source = &dto.ClaudeMessageSource{
 							Type: "base64",
 						}
 						// 判断是否是url
@@ -272,7 +274,7 @@ func RequestOpenAI2ClaudeMessage(textRequest dto.GeneralOpenAIRequest) (*ClaudeR
 							common.SysError("tool call function arguments is not a map[string]any: " + fmt.Sprintf("%v", toolCall.Function.Arguments))
 							continue
 						}
-						claudeMediaMessages = append(claudeMediaMessages, ClaudeMediaMessage{
+						claudeMediaMessages = append(claudeMediaMessages, dto.ClaudeMediaMessage{
 							Type:  "tool_use",
 							Id:    toolCall.ID,
 							Name:  toolCall.Function.Name,
@@ -290,13 +292,19 @@ func RequestOpenAI2ClaudeMessage(textRequest dto.GeneralOpenAIRequest) (*ClaudeR
 	return &claudeRequest, nil
 }
 
-func StreamResponseClaude2OpenAI(reqMode int, claudeResponse *ClaudeResponse) (*dto.ChatCompletionsStreamResponse, *ClaudeUsage) {
+func StreamResponseClaude2OpenAI(reqMode int, claudeResponse *dto.ClaudeResponse) *dto.ChatCompletionsStreamResponse {
 	var response dto.ChatCompletionsStreamResponse
-	var claudeUsage *ClaudeUsage
 	response.Object = "chat.completion.chunk"
 	response.Model = claudeResponse.Model
 	response.Choices = make([]dto.ChatCompletionsStreamResponseChoice, 0)
 	tools := make([]dto.ToolCallResponse, 0)
+	fcIdx := 0
+	if claudeResponse.Index != nil {
+		fcIdx = *claudeResponse.Index - 1
+		if fcIdx < 0 {
+			fcIdx = 0
+		}
+	}
 	var choice dto.ChatCompletionsStreamResponseChoice
 	if reqMode == RequestModeCompletion {
 		choice.Delta.SetContentString(claudeResponse.Completion)
@@ -308,7 +316,7 @@ func StreamResponseClaude2OpenAI(reqMode int, claudeResponse *ClaudeResponse) (*
 		if claudeResponse.Type == "message_start" {
 			response.Id = claudeResponse.Message.Id
 			response.Model = claudeResponse.Message.Model
-			claudeUsage = &claudeResponse.Message.Usage
+			//claudeUsage = &claudeResponse.Message.Usage
 			choice.Delta.SetContentString("")
 			choice.Delta.Role = "assistant"
 		} else if claudeResponse.Type == "content_block_start" {
@@ -316,8 +324,9 @@ func StreamResponseClaude2OpenAI(reqMode int, claudeResponse *ClaudeResponse) (*
 				//choice.Delta.SetContentString(claudeResponse.ContentBlock.Text)
 				if claudeResponse.ContentBlock.Type == "tool_use" {
 					tools = append(tools, dto.ToolCallResponse{
-						ID:   claudeResponse.ContentBlock.Id,
-						Type: "function",
+						Index: common.GetPointer(fcIdx),
+						ID:    claudeResponse.ContentBlock.Id,
+						Type:  "function",
 						Function: dto.FunctionResponse{
 							Name:      claudeResponse.ContentBlock.Name,
 							Arguments: "",
@@ -325,17 +334,18 @@ func StreamResponseClaude2OpenAI(reqMode int, claudeResponse *ClaudeResponse) (*
 					})
 				}
 			} else {
-				return nil, nil
+				return nil
 			}
 		} else if claudeResponse.Type == "content_block_delta" {
 			if claudeResponse.Delta != nil {
-				choice.Index = claudeResponse.Index
-				choice.Delta.SetContentString(claudeResponse.Delta.Text)
+				choice.Delta.Content = claudeResponse.Delta.Text
 				switch claudeResponse.Delta.Type {
 				case "input_json_delta":
 					tools = append(tools, dto.ToolCallResponse{
+						Type:  "function",
+						Index: common.GetPointer(fcIdx),
 						Function: dto.FunctionResponse{
-							Arguments: claudeResponse.Delta.PartialJson,
+							Arguments: *claudeResponse.Delta.PartialJson,
 						},
 					})
 				case "signature_delta":
@@ -352,26 +362,23 @@ func StreamResponseClaude2OpenAI(reqMode int, claudeResponse *ClaudeResponse) (*
 			if finishReason != "null" {
 				choice.FinishReason = &finishReason
 			}
-			claudeUsage = &claudeResponse.Usage
+			//claudeUsage = &claudeResponse.Usage
 		} else if claudeResponse.Type == "message_stop" {
-			return nil, nil
+			return nil
 		} else {
-			return nil, nil
+			return nil
 		}
 	}
-	if claudeUsage == nil {
-		claudeUsage = &ClaudeUsage{}
-	}
 	if len(tools) > 0 {
 		choice.Delta.Content = nil // compatible with other OpenAI derivative applications, like LobeOpenAICompatibleFactory ...
 		choice.Delta.ToolCalls = tools
 	}
 	response.Choices = append(response.Choices, choice)
 
-	return &response, claudeUsage
+	return &response
 }
 
-func ResponseClaude2OpenAI(reqMode int, claudeResponse *ClaudeResponse) *dto.OpenAITextResponse {
+func ResponseClaude2OpenAI(reqMode int, claudeResponse *dto.ClaudeResponse) *dto.OpenAITextResponse {
 	choices := make([]dto.OpenAITextResponseChoice, 0)
 	fullTextResponse := dto.OpenAITextResponse{
 		Id:      fmt.Sprintf("chatcmpl-%s", common.GetUUID()),
@@ -379,19 +386,20 @@ func ResponseClaude2OpenAI(reqMode int, claudeResponse *ClaudeResponse) *dto.Ope
 		Created: common.GetTimestamp(),
 	}
 	var responseText string
+	var responseThinking string
 	if len(claudeResponse.Content) > 0 {
-		responseText = claudeResponse.Content[0].Text
+		responseText = claudeResponse.Content[0].GetText()
+		responseThinking = claudeResponse.Content[0].Thinking
 	}
 	tools := make([]dto.ToolCallResponse, 0)
 	thinkingContent := ""
 
 	if reqMode == RequestModeCompletion {
-		content, _ := json.Marshal(strings.TrimPrefix(claudeResponse.Completion, " "))
 		choice := dto.OpenAITextResponseChoice{
 			Index: 0,
 			Message: dto.Message{
 				Role:    "assistant",
-				Content: content,
+				Content: strings.TrimPrefix(claudeResponse.Completion, " "),
 				Name:    nil,
 			},
 			FinishReason: stopReasonClaude2OpenAI(claudeResponse.StopReason),
@@ -415,7 +423,7 @@ func ResponseClaude2OpenAI(reqMode int, claudeResponse *ClaudeResponse) *dto.Ope
 				// 加密的不管， 只输出明文的推理过程
 				thinkingContent = message.Thinking
 			case "text":
-				responseText = message.Text
+				responseText = message.GetText()
 			}
 		}
 	}
@@ -427,6 +435,9 @@ func ResponseClaude2OpenAI(reqMode int, claudeResponse *ClaudeResponse) *dto.Ope
 		FinishReason: stopReasonClaude2OpenAI(claudeResponse.StopReason),
 	}
 	choice.SetStringContent(responseText)
+	if len(responseThinking) > 0 {
+		choice.ReasoningContent = responseThinking
+	}
 	if len(tools) > 0 {
 		choice.Message.SetToolCalls(tools)
 	}
@@ -437,126 +448,228 @@ func ResponseClaude2OpenAI(reqMode int, claudeResponse *ClaudeResponse) *dto.Ope
 	return &fullTextResponse
 }
 
-func ClaudeStreamHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo, requestMode int) (*dto.OpenAIErrorWithStatusCode, *dto.Usage) {
-	responseId := fmt.Sprintf("chatcmpl-%s", common.GetUUID())
-	var usage *dto.Usage
-	usage = &dto.Usage{}
-	responseText := ""
-	createdTime := common.GetTimestamp()
+type ClaudeResponseInfo struct {
+	ResponseId   string
+	Created      int64
+	Model        string
+	ResponseText strings.Builder
+	Usage        *dto.Usage
+}
 
-	helper.StreamScannerHandler(c, resp, info, func(data string) bool {
-		var claudeResponse ClaudeResponse
-		err := json.Unmarshal([]byte(data), &claudeResponse)
-		if err != nil {
-			common.SysError("error unmarshalling stream response: " + err.Error())
-			return true
+func FormatClaudeResponseInfo(requestMode int, claudeResponse *dto.ClaudeResponse, oaiResponse *dto.ChatCompletionsStreamResponse, claudeInfo *ClaudeResponseInfo) bool {
+	if requestMode == RequestModeCompletion {
+		claudeInfo.ResponseText.WriteString(claudeResponse.Completion)
+	} else {
+		if claudeResponse.Type == "message_start" {
+			// message_start, 获取usage
+			claudeInfo.ResponseId = claudeResponse.Message.Id
+			claudeInfo.Model = claudeResponse.Message.Model
+			claudeInfo.Usage.PromptTokens = claudeResponse.Message.Usage.InputTokens
+		} else if claudeResponse.Type == "content_block_delta" {
+			if claudeResponse.Delta.Text != nil {
+				claudeInfo.ResponseText.WriteString(*claudeResponse.Delta.Text)
+			}
+		} else if claudeResponse.Type == "message_delta" {
+			claudeInfo.Usage.CompletionTokens = claudeResponse.Usage.OutputTokens
+			if claudeResponse.Usage.InputTokens > 0 {
+				claudeInfo.Usage.PromptTokens = claudeResponse.Usage.InputTokens
+			}
+			claudeInfo.Usage.TotalTokens = claudeInfo.Usage.PromptTokens + claudeResponse.Usage.OutputTokens
+		} else if claudeResponse.Type == "content_block_start" {
+		} else {
+			return false
 		}
+	}
+	if oaiResponse != nil {
+		oaiResponse.Id = claudeInfo.ResponseId
+		oaiResponse.Created = claudeInfo.Created
+		oaiResponse.Model = claudeInfo.Model
+	}
+	return true
+}
 
-		response, claudeUsage := StreamResponseClaude2OpenAI(requestMode, &claudeResponse)
-		if response == nil {
-			return true
+func HandleStreamResponseData(c *gin.Context, info *relaycommon.RelayInfo, claudeInfo *ClaudeResponseInfo, data string, requestMode int) *dto.OpenAIErrorWithStatusCode {
+	var claudeResponse dto.ClaudeResponse
+	err := common.DecodeJsonStr(data, &claudeResponse)
+	if err != nil {
+		common.SysError("error unmarshalling stream response: " + err.Error())
+		return service.OpenAIErrorWrapper(err, "stream_response_error", http.StatusInternalServerError)
+	}
+	if claudeResponse.Error != nil && claudeResponse.Error.Type != "" {
+		return &dto.OpenAIErrorWithStatusCode{
+			Error: dto.OpenAIError{
+				Code:    "stream_response_error",
+				Type:    claudeResponse.Error.Type,
+				Message: claudeResponse.Error.Message,
+			},
+			StatusCode: http.StatusInternalServerError,
 		}
+	}
+	if info.RelayFormat == relaycommon.RelayFormatClaude {
 		if requestMode == RequestModeCompletion {
-			responseText += claudeResponse.Completion
-			responseId = response.Id
+			claudeInfo.ResponseText.WriteString(claudeResponse.Completion)
 		} else {
 			if claudeResponse.Type == "message_start" {
 				// message_start, 获取usage
-				responseId = claudeResponse.Message.Id
 				info.UpstreamModelName = claudeResponse.Message.Model
-				usage.PromptTokens = claudeUsage.InputTokens
+				claudeInfo.Usage.PromptTokens = claudeResponse.Message.Usage.InputTokens
+				claudeInfo.Usage.PromptTokensDetails.CachedTokens = claudeResponse.Message.Usage.CacheReadInputTokens
+				claudeInfo.Usage.PromptTokensDetails.CachedCreationTokens = claudeResponse.Message.Usage.CacheCreationInputTokens
+				claudeInfo.Usage.CompletionTokens = claudeResponse.Message.Usage.OutputTokens
 			} else if claudeResponse.Type == "content_block_delta" {
-				responseText += claudeResponse.Delta.Text
+				claudeInfo.ResponseText.WriteString(claudeResponse.Delta.GetText())
 			} else if claudeResponse.Type == "message_delta" {
-				usage.CompletionTokens = claudeUsage.OutputTokens
-				usage.TotalTokens = claudeUsage.InputTokens + claudeUsage.OutputTokens
-			} else if claudeResponse.Type == "content_block_start" {
-				return true
-			} else {
-				return true
+				if claudeResponse.Usage.InputTokens > 0 {
+					// 不叠加，只取最新的
+					claudeInfo.Usage.PromptTokens = claudeResponse.Usage.InputTokens
+				}
+				claudeInfo.Usage.CompletionTokens = claudeResponse.Usage.OutputTokens
+				claudeInfo.Usage.TotalTokens = claudeInfo.Usage.PromptTokens + claudeInfo.Usage.CompletionTokens
 			}
 		}
-		//response.Id = responseId
-		response.Id = responseId
-		response.Created = createdTime
-		response.Model = info.UpstreamModelName
+		helper.ClaudeChunkData(c, claudeResponse, data)
+	} else if info.RelayFormat == relaycommon.RelayFormatOpenAI {
+		response := StreamResponseClaude2OpenAI(requestMode, &claudeResponse)
+
+		if !FormatClaudeResponseInfo(requestMode, &claudeResponse, response, claudeInfo) {
+			return nil
+		}
 
 		err = helper.ObjectData(c, response)
 		if err != nil {
 			common.LogError(c, "send_stream_response_failed: "+err.Error())
 		}
-		return true
-	})
-
-	if requestMode == RequestModeCompletion {
-		usage, _ = service.ResponseText2Usage(responseText, info.UpstreamModelName, info.PromptTokens)
-	} else {
-		if usage.PromptTokens == 0 {
-			usage.PromptTokens = info.PromptTokens
-		}
-		if usage.CompletionTokens == 0 {
-			usage, _ = service.ResponseText2Usage(responseText, info.UpstreamModelName, usage.PromptTokens)
-		}
 	}
-	if info.ShouldIncludeUsage {
-		response := helper.GenerateFinalUsageResponse(responseId, createdTime, info.UpstreamModelName, *usage)
-		err := helper.ObjectData(c, response)
-		if err != nil {
-			common.SysError("send final response failed: " + err.Error())
-		}
-	}
-	helper.Done(c)
-	//resp.Body.Close()
-	return nil, usage
+	return nil
 }
 
-func ClaudeHandler(c *gin.Context, resp *http.Response, requestMode int, info *relaycommon.RelayInfo) (*dto.OpenAIErrorWithStatusCode, *dto.Usage) {
-	responseBody, err := io.ReadAll(resp.Body)
-	if err != nil {
-		return service.OpenAIErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
+func HandleStreamFinalResponse(c *gin.Context, info *relaycommon.RelayInfo, claudeInfo *ClaudeResponseInfo, requestMode int) {
+	if info.RelayFormat == relaycommon.RelayFormatClaude {
+		if requestMode == RequestModeCompletion {
+			claudeInfo.Usage, _ = service.ResponseText2Usage(claudeInfo.ResponseText.String(), info.UpstreamModelName, info.PromptTokens)
+		} else {
+			// 说明流模式建立失败，可能为官方出错
+			if claudeInfo.Usage.PromptTokens == 0 {
+				//usage.PromptTokens = info.PromptTokens
+			}
+			if claudeInfo.Usage.CompletionTokens == 0 {
+				claudeInfo.Usage, _ = service.ResponseText2Usage(claudeInfo.ResponseText.String(), info.UpstreamModelName, claudeInfo.Usage.PromptTokens)
+			}
+		}
+	} else if info.RelayFormat == relaycommon.RelayFormatOpenAI {
+		if requestMode == RequestModeCompletion {
+			claudeInfo.Usage, _ = service.ResponseText2Usage(claudeInfo.ResponseText.String(), info.UpstreamModelName, info.PromptTokens)
+		} else {
+			if claudeInfo.Usage.PromptTokens == 0 {
+				//上游出错
+			}
+			if claudeInfo.Usage.CompletionTokens == 0 {
+				claudeInfo.Usage, _ = service.ResponseText2Usage(claudeInfo.ResponseText.String(), info.UpstreamModelName, claudeInfo.Usage.PromptTokens)
+			}
+		}
+		if info.ShouldIncludeUsage {
+			response := helper.GenerateFinalUsageResponse(claudeInfo.ResponseId, claudeInfo.Created, info.UpstreamModelName, *claudeInfo.Usage)
+			err := helper.ObjectData(c, response)
+			if err != nil {
+				common.SysError("send final response failed: " + err.Error())
+			}
+		}
+		helper.Done(c)
 	}
-	err = resp.Body.Close()
-	if err != nil {
-		return service.OpenAIErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+}
+
+func ClaudeStreamHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo, requestMode int) (*dto.OpenAIErrorWithStatusCode, *dto.Usage) {
+	claudeInfo := &ClaudeResponseInfo{
+		ResponseId:   fmt.Sprintf("chatcmpl-%s", common.GetUUID()),
+		Created:      common.GetTimestamp(),
+		Model:        info.UpstreamModelName,
+		ResponseText: strings.Builder{},
+		Usage:        &dto.Usage{},
 	}
-	var claudeResponse ClaudeResponse
-	err = json.Unmarshal(responseBody, &claudeResponse)
+	var err *dto.OpenAIErrorWithStatusCode
+	helper.StreamScannerHandler(c, resp, info, func(data string) bool {
+		err = HandleStreamResponseData(c, info, claudeInfo, data, requestMode)
+		if err != nil {
+			return false
+		}
+		return true
+	})
 	if err != nil {
-		return service.OpenAIErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
+		return err, nil
 	}
-	if claudeResponse.Error.Type != "" {
+
+	HandleStreamFinalResponse(c, info, claudeInfo, requestMode)
+	return nil, claudeInfo.Usage
+}
+
+func HandleClaudeResponseData(c *gin.Context, info *relaycommon.RelayInfo, claudeInfo *ClaudeResponseInfo, data []byte, requestMode int) *dto.OpenAIErrorWithStatusCode {
+	var claudeResponse dto.ClaudeResponse
+	err := common.DecodeJson(data, &claudeResponse)
+	if err != nil {
+		return service.OpenAIErrorWrapper(err, "unmarshal_claude_response_failed", http.StatusInternalServerError)
+	}
+	if claudeResponse.Error != nil && claudeResponse.Error.Type != "" {
 		return &dto.OpenAIErrorWithStatusCode{
 			Error: dto.OpenAIError{
 				Message: claudeResponse.Error.Message,
 				Type:    claudeResponse.Error.Type,
-				Param:   "",
 				Code:    claudeResponse.Error.Type,
 			},
-			StatusCode: resp.StatusCode,
-		}, nil
+			StatusCode: http.StatusInternalServerError,
+		}
 	}
-	fullTextResponse := ResponseClaude2OpenAI(requestMode, &claudeResponse)
-	completionTokens, err := service.CountTextToken(claudeResponse.Completion, info.OriginModelName)
-	if err != nil {
-		return service.OpenAIErrorWrapper(err, "count_token_text_failed", http.StatusInternalServerError), nil
-	}
-	usage := dto.Usage{}
 	if requestMode == RequestModeCompletion {
-		usage.PromptTokens = info.PromptTokens
-		usage.CompletionTokens = completionTokens
-		usage.TotalTokens = info.PromptTokens + completionTokens
+		completionTokens, err := service.CountTextToken(claudeResponse.Completion, info.OriginModelName)
+		if err != nil {
+			return service.OpenAIErrorWrapper(err, "count_token_text_failed", http.StatusInternalServerError)
+		}
+		claudeInfo.Usage.PromptTokens = info.PromptTokens
+		claudeInfo.Usage.CompletionTokens = completionTokens
+		claudeInfo.Usage.TotalTokens = info.PromptTokens + completionTokens
 	} else {
-		usage.PromptTokens = claudeResponse.Usage.InputTokens
-		usage.CompletionTokens = claudeResponse.Usage.OutputTokens
-		usage.TotalTokens = claudeResponse.Usage.InputTokens + claudeResponse.Usage.OutputTokens
+		claudeInfo.Usage.PromptTokens = claudeResponse.Usage.InputTokens
+		claudeInfo.Usage.CompletionTokens = claudeResponse.Usage.OutputTokens
+		claudeInfo.Usage.TotalTokens = claudeResponse.Usage.InputTokens + claudeResponse.Usage.OutputTokens
+		claudeInfo.Usage.PromptTokensDetails.CachedTokens = claudeResponse.Usage.CacheReadInputTokens
+		claudeInfo.Usage.PromptTokensDetails.CachedCreationTokens = claudeResponse.Usage.CacheCreationInputTokens
 	}
-	fullTextResponse.Usage = usage
-	jsonResponse, err := json.Marshal(fullTextResponse)
-	if err != nil {
-		return service.OpenAIErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
+	var responseData []byte
+	switch info.RelayFormat {
+	case relaycommon.RelayFormatOpenAI:
+		openaiResponse := ResponseClaude2OpenAI(requestMode, &claudeResponse)
+		openaiResponse.Usage = *claudeInfo.Usage
+		responseData, err = json.Marshal(openaiResponse)
+		if err != nil {
+			return service.OpenAIErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError)
+		}
+	case relaycommon.RelayFormatClaude:
+		responseData = data
 	}
 	c.Writer.Header().Set("Content-Type", "application/json")
-	c.Writer.WriteHeader(resp.StatusCode)
-	_, err = c.Writer.Write(jsonResponse)
-	return nil, &usage
+	c.Writer.WriteHeader(http.StatusOK)
+	_, err = c.Writer.Write(responseData)
+	return nil
+}
+
+func ClaudeHandler(c *gin.Context, resp *http.Response, requestMode int, info *relaycommon.RelayInfo) (*dto.OpenAIErrorWithStatusCode, *dto.Usage) {
+	claudeInfo := &ClaudeResponseInfo{
+		ResponseId:   fmt.Sprintf("chatcmpl-%s", common.GetUUID()),
+		Created:      common.GetTimestamp(),
+		Model:        info.UpstreamModelName,
+		ResponseText: strings.Builder{},
+		Usage:        &dto.Usage{},
+	}
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return service.OpenAIErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
+	}
+	resp.Body.Close()
+	if common.DebugEnabled {
+		println("responseBody: ", string(responseBody))
+	}
+	handleErr := HandleClaudeResponseData(c, info, claudeInfo, responseBody, requestMode)
+	if handleErr != nil {
+		return handleErr, nil
+	}
+	return nil, claudeInfo.Usage
 }

relay/channel/cloudflare/adaptor.go

@@ -17,6 +17,12 @@ import (
 type Adaptor struct {
 }
 
+func (a *Adaptor) ConvertClaudeRequest(*gin.Context, *relaycommon.RelayInfo, *dto.ClaudeRequest) (any, error) {
+	//TODO implement me
+	panic("implement me")
+	return nil, nil
+}
+
 func (a *Adaptor) Init(info *relaycommon.RelayInfo) {
 }
 
@@ -37,7 +43,7 @@ func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *rel
 	return nil
 }
 
-func (a *Adaptor) ConvertRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
+func (a *Adaptor) ConvertOpenAIRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
 	if request == nil {
 		return nil, errors.New("request is nil")
 	}
@@ -49,6 +55,11 @@ func (a *Adaptor) ConvertRequest(c *gin.Context, info *relaycommon.RelayInfo, re
 	}
 }
 
+func (a *Adaptor) ConvertOpenAIResponsesRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.OpenAIResponsesRequest) (any, error) {
+	// TODO implement me
+	return nil, errors.New("not implemented")
+}
+
 func (a *Adaptor) DoRequest(c *gin.Context, info *relaycommon.RelayInfo, requestBody io.Reader) (any, error) {
 	return channel.DoApiRequest(a, c, info, requestBody)
 }

relay/channel/cohere/adaptor.go

@@ -3,18 +3,25 @@ package cohere
 import (
 	"errors"
 	"fmt"
-	"github.com/gin-gonic/gin"
 	"io"
 	"net/http"
 	"one-api/dto"
 	"one-api/relay/channel"
 	relaycommon "one-api/relay/common"
 	"one-api/relay/constant"
+
+	"github.com/gin-gonic/gin"
 )
 
 type Adaptor struct {
 }
 
+func (a *Adaptor) ConvertClaudeRequest(*gin.Context, *relaycommon.RelayInfo, *dto.ClaudeRequest) (any, error) {
+	//TODO implement me
+	panic("implement me")
+	return nil, nil
+}
+
 func (a *Adaptor) ConvertAudioRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.AudioRequest) (io.Reader, error) {
 	//TODO implement me
 	return nil, errors.New("not implemented")
@@ -42,10 +49,15 @@ func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *rel
 	return nil
 }
 
-func (a *Adaptor) ConvertRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
+func (a *Adaptor) ConvertOpenAIRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
 	return requestOpenAI2Cohere(*request), nil
 }
 
+func (a *Adaptor) ConvertOpenAIResponsesRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.OpenAIResponsesRequest) (any, error) {
+	// TODO implement me
+	return nil, errors.New("not implemented")
+}
+
 func (a *Adaptor) DoRequest(c *gin.Context, info *relaycommon.RelayInfo, requestBody io.Reader) (any, error) {
 	return channel.DoApiRequest(a, c, info, requestBody)
 }
@@ -59,7 +71,6 @@ func (a *Adaptor) ConvertEmbeddingRequest(c *gin.Context, info *relaycommon.Rela
 	return nil, errors.New("not implemented")
 }
 
-
 func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (usage any, err *dto.OpenAIErrorWithStatusCode) {
 	if info.RelayMode == constant.RelayModeRerank {
 		err, usage = cohereRerankHandler(c, resp, info)

relay/channel/cohere/constant.go

@@ -1,6 +1,7 @@
 package cohere
 
 var ModelList = []string{
+	"command-a-03-2025",
 	"command-r", "command-r-plus",
 	"command-r-08-2024", "command-r-plus-08-2024",
 	"c4ai-aya-23-35b", "c4ai-aya-23-8b",

relay/channel/cohere/dto.go

@@ -40,8 +40,8 @@ type CohereRerankRequest struct {
 }
 
 type CohereRerankResponseResult struct {
-	Results []dto.RerankResponseDocument `json:"results"`
-	Meta    CohereMeta                   `json:"meta"`
+	Results []dto.RerankResponseResult `json:"results"`
+	Meta    CohereMeta                 `json:"meta"`
 }
 
 type CohereMeta struct {

relay/channel/cohere/relay-cohere.go

@@ -195,11 +195,10 @@ func cohereHandler(c *gin.Context, resp *http.Response, modelName string, prompt
 	openaiResp.Model = modelName
 	openaiResp.Usage = usage
 
-	content, _ := json.Marshal(cohereResp.Text)
 	openaiResp.Choices = []dto.OpenAITextResponseChoice{
 		{
 			Index:        0,
-			Message:      dto.Message{Content: content, Role: "assistant"},
+			Message:      dto.Message{Content: cohereResp.Text, Role: "assistant"},
 			FinishReason: stopReasonCohere2OpenAI(cohereResp.FinishReason),
 		},
 	}

relay/channel/coze/adaptor.go

@@ -0,0 +1,132 @@
+package coze
+
+import (
+	"encoding/json"
+	"errors"
+	"fmt"
+	"io"
+	"net/http"
+	"one-api/dto"
+	"one-api/relay/channel"
+	"one-api/relay/common"
+	"time"
+
+	"github.com/gin-gonic/gin"
+)
+
+type Adaptor struct {
+}
+
+// ConvertAudioRequest implements channel.Adaptor.
+func (a *Adaptor) ConvertAudioRequest(c *gin.Context, info *common.RelayInfo, request dto.AudioRequest) (io.Reader, error) {
+	return nil, errors.New("not implemented")
+}
+
+// ConvertClaudeRequest implements channel.Adaptor.
+func (a *Adaptor) ConvertClaudeRequest(c *gin.Context, info *common.RelayInfo, request *dto.ClaudeRequest) (any, error) {
+	return nil, errors.New("not implemented")
+}
+
+// ConvertEmbeddingRequest implements channel.Adaptor.
+func (a *Adaptor) ConvertEmbeddingRequest(c *gin.Context, info *common.RelayInfo, request dto.EmbeddingRequest) (any, error) {
+	return nil, errors.New("not implemented")
+}
+
+// ConvertImageRequest implements channel.Adaptor.
+func (a *Adaptor) ConvertImageRequest(c *gin.Context, info *common.RelayInfo, request dto.ImageRequest) (any, error) {
+	return nil, errors.New("not implemented")
+}
+
+// ConvertOpenAIRequest implements channel.Adaptor.
+func (a *Adaptor) ConvertOpenAIRequest(c *gin.Context, info *common.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
+	if request == nil {
+		return nil, errors.New("request is nil")
+	}
+	return convertCozeChatRequest(c, *request), nil
+}
+
+// ConvertOpenAIResponsesRequest implements channel.Adaptor.
+func (a *Adaptor) ConvertOpenAIResponsesRequest(c *gin.Context, info *common.RelayInfo, request dto.OpenAIResponsesRequest) (any, error) {
+	return nil, errors.New("not implemented")
+}
+
+// ConvertRerankRequest implements channel.Adaptor.
+func (a *Adaptor) ConvertRerankRequest(c *gin.Context, relayMode int, request dto.RerankRequest) (any, error) {
+	return nil, errors.New("not implemented")
+}
+
+// DoRequest implements channel.Adaptor.
+func (a *Adaptor) DoRequest(c *gin.Context, info *common.RelayInfo, requestBody io.Reader) (any, error) {
+	if info.IsStream {
+		return channel.DoApiRequest(a, c, info, requestBody)
+	}
+	// 首先发送创建消息请求，成功后再发送获取消息请求
+	// 发送创建消息请求
+	resp, err := channel.DoApiRequest(a, c, info, requestBody)
+	if err != nil {
+		return nil, err
+	}
+	// 解析 resp
+	var cozeResponse CozeChatResponse
+	respBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return nil, err
+	}
+	err = json.Unmarshal(respBody, &cozeResponse)
+	if cozeResponse.Code != 0 {
+		return nil, errors.New(cozeResponse.Msg)
+	}
+	c.Set("coze_conversation_id", cozeResponse.Data.ConversationId)
+	c.Set("coze_chat_id", cozeResponse.Data.Id)
+	// 轮询检查消息是否完成
+	for {
+		err, isComplete := checkIfChatComplete(a, c, info)
+		if err != nil {
+			return nil, err
+		} else {
+			if isComplete {
+				break
+			}
+		}
+		time.Sleep(time.Second * 1)
+	}
+	// 发送获取消息请求
+	return getChatDetail(a, c, info)
+}
+
+// DoResponse implements channel.Adaptor.
+func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, info *common.RelayInfo) (usage any, err *dto.OpenAIErrorWithStatusCode) {
+	if info.IsStream {
+		err, usage = cozeChatStreamHandler(c, resp, info)
+	} else {
+		err, usage = cozeChatHandler(c, resp, info)
+	}
+	return
+}
+
+// GetChannelName implements channel.Adaptor.
+func (a *Adaptor) GetChannelName() string {
+	return ChannelName
+}
+
+// GetModelList implements channel.Adaptor.
+func (a *Adaptor) GetModelList() []string {
+	return ModelList
+}
+
+// GetRequestURL implements channel.Adaptor.
+func (a *Adaptor) GetRequestURL(info *common.RelayInfo) (string, error) {
+	return fmt.Sprintf("%s/v3/chat", info.BaseUrl), nil
+}
+
+// Init implements channel.Adaptor.
+func (a *Adaptor) Init(info *common.RelayInfo) {
+
+}
+
+// SetupRequestHeader implements channel.Adaptor.
+func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *common.RelayInfo) error {
+	channel.SetupApiRequestHeader(info, c, req)
+	req.Set("Authorization", "Bearer "+info.ApiKey)
+	return nil
+}

relay/channel/coze/constants.go

@@ -0,0 +1,30 @@
+package coze
+
+var ModelList = []string{
+	"moonshot-v1-8k",
+	"moonshot-v1-32k",
+	"moonshot-v1-128k",
+	"Baichuan4",
+	"abab6.5s-chat-pro",
+	"glm-4-0520",
+	"qwen-max",
+	"deepseek-r1",
+	"deepseek-v3",
+	"deepseek-r1-distill-qwen-32b",
+	"deepseek-r1-distill-qwen-7b",
+	"step-1v-8k",
+	"step-1.5v-mini",
+	"Doubao-pro-32k",
+	"Doubao-pro-256k",
+	"Doubao-lite-128k",
+	"Doubao-lite-32k",
+	"Doubao-vision-lite-32k",
+	"Doubao-vision-pro-32k",
+	"Doubao-1.5-pro-vision-32k",
+	"Doubao-1.5-lite-32k",
+	"Doubao-1.5-pro-32k",
+	"Doubao-1.5-thinking-pro",
+	"Doubao-1.5-pro-256k",
+}
+
+var ChannelName = "coze"

relay/channel/coze/dto.go

@@ -0,0 +1,78 @@
+package coze
+
+import "encoding/json"
+
+type CozeError struct {
+	Code    int    `json:"code"`
+	Message string `json:"message"`
+}
+
+type CozeEnterMessage struct {
+	Role        string          `json:"role"`
+	Type        string          `json:"type,omitempty"`
+	Content     any             `json:"content,omitempty"`
+	MetaData    json.RawMessage `json:"meta_data,omitempty"`
+	ContentType string          `json:"content_type,omitempty"`
+}
+
+type CozeChatRequest struct {
+	BotId              string             `json:"bot_id"`
+	UserId             string             `json:"user_id"`
+	AdditionalMessages []CozeEnterMessage `json:"additional_messages,omitempty"`
+	Stream             bool               `json:"stream,omitempty"`
+	CustomVariables    json.RawMessage    `json:"custom_variables,omitempty"`
+	AutoSaveHistory    bool               `json:"auto_save_history,omitempty"`
+	MetaData           json.RawMessage    `json:"meta_data,omitempty"`
+	ExtraParams        json.RawMessage    `json:"extra_params,omitempty"`
+	ShortcutCommand    json.RawMessage    `json:"shortcut_command,omitempty"`
+	Parameters         json.RawMessage    `json:"parameters,omitempty"`
+}
+
+type CozeChatResponse struct {
+	Code int                  `json:"code"`
+	Msg  string               `json:"msg"`
+	Data CozeChatResponseData `json:"data"`
+}
+
+type CozeChatResponseData struct {
+	Id             string        `json:"id"`
+	ConversationId string        `json:"conversation_id"`
+	BotId          string        `json:"bot_id"`
+	CreatedAt      int64         `json:"created_at"`
+	LastError      CozeError     `json:"last_error"`
+	Status         string        `json:"status"`
+	Usage          CozeChatUsage `json:"usage"`
+}
+
+type CozeChatUsage struct {
+	TokenCount  int `json:"token_count"`
+	OutputCount int `json:"output_count"`
+	InputCount  int `json:"input_count"`
+}
+
+type CozeChatDetailResponse struct {
+	Data   []CozeChatV3MessageDetail `json:"data"`
+	Code   int                       `json:"code"`
+	Msg    string                    `json:"msg"`
+	Detail CozeResponseDetail        `json:"detail"`
+}
+
+type CozeChatV3MessageDetail struct {
+	Id               string          `json:"id"`
+	Role             string          `json:"role"`
+	Type             string          `json:"type"`
+	BotId            string          `json:"bot_id"`
+	ChatId           string          `json:"chat_id"`
+	Content          json.RawMessage `json:"content"`
+	MetaData         json.RawMessage `json:"meta_data"`
+	CreatedAt        int64           `json:"created_at"`
+	SectionId        string          `json:"section_id"`
+	UpdatedAt        int64           `json:"updated_at"`
+	ContentType      string          `json:"content_type"`
+	ConversationId   string          `json:"conversation_id"`
+	ReasoningContent string          `json:"reasoning_content"`
+}
+
+type CozeResponseDetail struct {
+	Logid string `json:"logid"`
+}

relay/channel/coze/relay-coze.go

@@ -0,0 +1,300 @@
+package coze
+
+import (
+	"bufio"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"io"
+	"net/http"
+	"one-api/common"
+	"one-api/dto"
+	relaycommon "one-api/relay/common"
+	"one-api/relay/helper"
+	"one-api/service"
+	"strings"
+
+	"github.com/gin-gonic/gin"
+)
+
+func convertCozeChatRequest(c *gin.Context, request dto.GeneralOpenAIRequest) *CozeChatRequest {
+	var messages []CozeEnterMessage
+	// 将 request的messages的role为user的content转换为CozeMessage
+	for _, message := range request.Messages {
+		if message.Role == "user" {
+			messages = append(messages, CozeEnterMessage{
+				Role:    "user",
+				Content: message.Content,
+				// TODO: support more content type
+				ContentType: "text",
+			})
+		}
+	}
+	user := request.User
+	if user == "" {
+		user = helper.GetResponseID(c)
+	}
+	cozeRequest := &CozeChatRequest{
+		BotId:              c.GetString("bot_id"),
+		UserId:             user,
+		AdditionalMessages: messages,
+		Stream:             request.Stream,
+	}
+	return cozeRequest
+}
+
+func cozeChatHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (*dto.OpenAIErrorWithStatusCode, *dto.Usage) {
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return service.OpenAIErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError), nil
+	}
+	err = resp.Body.Close()
+	if err != nil {
+		return service.OpenAIErrorWrapperLocal(err, "close_response_body_failed", http.StatusInternalServerError), nil
+	}
+	// convert coze response to openai response
+	var response dto.TextResponse
+	var cozeResponse CozeChatDetailResponse
+	response.Model = info.UpstreamModelName
+	err = json.Unmarshal(responseBody, &cozeResponse)
+	if err != nil {
+		return service.OpenAIErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	if cozeResponse.Code != 0 {
+		return service.OpenAIErrorWrapper(errors.New(cozeResponse.Msg), fmt.Sprintf("%d", cozeResponse.Code), http.StatusInternalServerError), nil
+	}
+	// 从上下文获取 usage
+	var usage dto.Usage
+	usage.PromptTokens = c.GetInt("coze_input_count")
+	usage.CompletionTokens = c.GetInt("coze_output_count")
+	usage.TotalTokens = c.GetInt("coze_token_count")
+	response.Usage = usage
+	response.Id = helper.GetResponseID(c)
+
+	var responseContent json.RawMessage
+	for _, data := range cozeResponse.Data {
+		if data.Type == "answer" {
+			responseContent = data.Content
+			response.Created = data.CreatedAt
+		}
+	}
+	// 添加 response.Choices
+	response.Choices = []dto.OpenAITextResponseChoice{
+		{
+			Index:        0,
+			Message:      dto.Message{Role: "assistant", Content: responseContent},
+			FinishReason: "stop",
+		},
+	}
+	jsonResponse, err := json.Marshal(response)
+	if err != nil {
+		return service.OpenAIErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil
+	}
+	c.Writer.Header().Set("Content-Type", "application/json")
+	c.Writer.WriteHeader(resp.StatusCode)
+	_, _ = c.Writer.Write(jsonResponse)
+
+	return nil, &usage
+}
+
+func cozeChatStreamHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (*dto.OpenAIErrorWithStatusCode, *dto.Usage) {
+	scanner := bufio.NewScanner(resp.Body)
+	scanner.Split(bufio.ScanLines)
+	helper.SetEventStreamHeaders(c)
+	id := helper.GetResponseID(c)
+	var responseText string
+
+	var currentEvent string
+	var currentData string
+	var usage dto.Usage
+
+	for scanner.Scan() {
+		line := scanner.Text()
+
+		if line == "" {
+			if currentEvent != "" && currentData != "" {
+				// handle last event
+				handleCozeEvent(c, currentEvent, currentData, &responseText, &usage, id, info)
+				currentEvent = ""
+				currentData = ""
+			}
+			continue
+		}
+
+		if strings.HasPrefix(line, "event:") {
+			currentEvent = strings.TrimSpace(line[6:])
+			continue
+		}
+
+		if strings.HasPrefix(line, "data:") {
+			currentData = strings.TrimSpace(line[5:])
+			continue
+		}
+	}
+
+	// Last event
+	if currentEvent != "" && currentData != "" {
+		handleCozeEvent(c, currentEvent, currentData, &responseText, &usage, id, info)
+	}
+
+	if err := scanner.Err(); err != nil {
+		return service.OpenAIErrorWrapper(err, "stream_scanner_error", http.StatusInternalServerError), nil
+	}
+	helper.Done(c)
+
+	if usage.TotalTokens == 0 {
+		usage.PromptTokens = info.PromptTokens
+		usage.CompletionTokens, _ = service.CountTextToken("gpt-3.5-turbo", responseText)
+		usage.TotalTokens = usage.PromptTokens + usage.CompletionTokens
+	}
+
+	return nil, &usage
+}
+
+func handleCozeEvent(c *gin.Context, event string, data string, responseText *string, usage *dto.Usage, id string, info *relaycommon.RelayInfo) {
+	switch event {
+	case "conversation.chat.completed":
+		// 将 data 解析为 CozeChatResponseData
+		var chatData CozeChatResponseData
+		err := json.Unmarshal([]byte(data), &chatData)
+		if err != nil {
+			common.SysError("error_unmarshalling_stream_response: " + err.Error())
+			return
+		}
+
+		usage.PromptTokens = chatData.Usage.InputCount
+		usage.CompletionTokens = chatData.Usage.OutputCount
+		usage.TotalTokens = chatData.Usage.TokenCount
+
+		finishReason := "stop"
+		stopResponse := helper.GenerateStopResponse(id, common.GetTimestamp(), info.UpstreamModelName, finishReason)
+		helper.ObjectData(c, stopResponse)
+
+	case "conversation.message.delta":
+		// 将 data 解析为 CozeChatV3MessageDetail
+		var messageData CozeChatV3MessageDetail
+		err := json.Unmarshal([]byte(data), &messageData)
+		if err != nil {
+			common.SysError("error_unmarshalling_stream_response: " + err.Error())
+			return
+		}
+
+		var content string
+		err = json.Unmarshal(messageData.Content, &content)
+		if err != nil {
+			common.SysError("error_unmarshalling_stream_response: " + err.Error())
+			return
+		}
+
+		*responseText += content
+
+		openaiResponse := dto.ChatCompletionsStreamResponse{
+			Id:      id,
+			Object:  "chat.completion.chunk",
+			Created: common.GetTimestamp(),
+			Model:   info.UpstreamModelName,
+		}
+
+		choice := dto.ChatCompletionsStreamResponseChoice{
+			Index: 0,
+		}
+		choice.Delta.SetContentString(content)
+		openaiResponse.Choices = append(openaiResponse.Choices, choice)
+
+		helper.ObjectData(c, openaiResponse)
+
+	case "error":
+		var errorData CozeError
+		err := json.Unmarshal([]byte(data), &errorData)
+		if err != nil {
+			common.SysError("error_unmarshalling_stream_response: " + err.Error())
+			return
+		}
+
+		common.SysError(fmt.Sprintf("stream event error: ", errorData.Code, errorData.Message))
+	}
+}
+
+func checkIfChatComplete(a *Adaptor, c *gin.Context, info *relaycommon.RelayInfo) (error, bool) {
+	requestURL := fmt.Sprintf("%s/v3/chat/retrieve", info.BaseUrl)
+
+	requestURL = requestURL + "?conversation_id=" + c.GetString("coze_conversation_id") + "&chat_id=" + c.GetString("coze_chat_id")
+	// 将 conversationId和chatId作为参数发送get请求
+	req, err := http.NewRequest("GET", requestURL, nil)
+	if err != nil {
+		return err, false
+	}
+	err = a.SetupRequestHeader(c, &req.Header, info)
+	if err != nil {
+		return err, false
+	}
+
+	resp, err := doRequest(req, info) // 调用 doRequest
+	if err != nil {
+		return err, false
+	}
+	if resp == nil { // 确保在 doRequest 失败时 resp 不为 nil 导致 panic
+		return fmt.Errorf("resp is nil"), false
+	}
+	defer resp.Body.Close() // 确保响应体被关闭
+
+	// 解析 resp 到 CozeChatResponse
+	var cozeResponse CozeChatResponse
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return fmt.Errorf("read response body failed: %w", err), false
+	}
+	err = json.Unmarshal(responseBody, &cozeResponse)
+	if err != nil {
+		return fmt.Errorf("unmarshal response body failed: %w", err), false
+	}
+	if cozeResponse.Data.Status == "completed" {
+		// 在上下文设置 usage
+		c.Set("coze_token_count", cozeResponse.Data.Usage.TokenCount)
+		c.Set("coze_output_count", cozeResponse.Data.Usage.OutputCount)
+		c.Set("coze_input_count", cozeResponse.Data.Usage.InputCount)
+		return nil, true
+	} else if cozeResponse.Data.Status == "failed" || cozeResponse.Data.Status == "canceled" || cozeResponse.Data.Status == "requires_action" {
+		return fmt.Errorf("chat status: %s", cozeResponse.Data.Status), false
+	} else {
+		return nil, false
+	}
+}
+
+func getChatDetail(a *Adaptor, c *gin.Context, info *relaycommon.RelayInfo) (*http.Response, error) {
+	requestURL := fmt.Sprintf("%s/v3/chat/message/list", info.BaseUrl)
+
+	requestURL = requestURL + "?conversation_id=" + c.GetString("coze_conversation_id") + "&chat_id=" + c.GetString("coze_chat_id")
+	req, err := http.NewRequest("GET", requestURL, nil)
+	if err != nil {
+		return nil, fmt.Errorf("new request failed: %w", err)
+	}
+	err = a.SetupRequestHeader(c, &req.Header, info)
+	if err != nil {
+		return nil, fmt.Errorf("setup request header failed: %w", err)
+	}
+	resp, err := doRequest(req, info)
+	if err != nil {
+		return nil, fmt.Errorf("do request failed: %w", err)
+	}
+	return resp, nil
+}
+
+func doRequest(req *http.Request, info *relaycommon.RelayInfo) (*http.Response, error) {
+	var client *http.Client
+	var err error // 声明 err 变量
+	if proxyURL, ok := info.ChannelSetting["proxy"]; ok {
+		client, err = service.NewProxyHttpClient(proxyURL.(string))
+		if err != nil {
+			return nil, fmt.Errorf("new proxy http client failed: %w", err)
+		}
+	} else {
+		client = service.GetHttpClient()
+	}
+	resp, err := client.Do(req)
+	if err != nil { // 增加对 client.Do(req) 返回错误的检查
+		return nil, fmt.Errorf("client.Do failed: %w", err)
+	}
+	// _ = resp.Body.Close()
+	return resp, nil
+}

relay/channel/deepseek/adaptor.go

@@ -3,7 +3,6 @@ package deepseek
 import (
 	"errors"
 	"fmt"
-	"github.com/gin-gonic/gin"
 	"io"
 	"net/http"
 	"one-api/dto"
@@ -11,11 +10,20 @@ import (
 	"one-api/relay/channel/openai"
 	relaycommon "one-api/relay/common"
 	"one-api/relay/constant"
+	"strings"
+
+	"github.com/gin-gonic/gin"
 )
 
 type Adaptor struct {
 }
 
+func (a *Adaptor) ConvertClaudeRequest(*gin.Context, *relaycommon.RelayInfo, *dto.ClaudeRequest) (any, error) {
+	//TODO implement me
+	panic("implement me")
+	return nil, nil
+}
+
 func (a *Adaptor) ConvertAudioRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.AudioRequest) (io.Reader, error) {
 	//TODO implement me
 	return nil, errors.New("not implemented")
@@ -30,9 +38,13 @@ func (a *Adaptor) Init(info *relaycommon.RelayInfo) {
 }
 
 func (a *Adaptor) GetRequestURL(info *relaycommon.RelayInfo) (string, error) {
+	fimBaseUrl := info.BaseUrl
+	if !strings.HasSuffix(info.BaseUrl, "/beta") {
+		fimBaseUrl += "/beta"
+	}
 	switch info.RelayMode {
 	case constant.RelayModeCompletions:
-		return fmt.Sprintf("%s/beta/completions", info.BaseUrl), nil
+		return fmt.Sprintf("%s/completions", fimBaseUrl), nil
 	default:
 		return fmt.Sprintf("%s/v1/chat/completions", info.BaseUrl), nil
 	}
@@ -44,7 +56,7 @@ func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *rel
 	return nil
 }
 
-func (a *Adaptor) ConvertRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
+func (a *Adaptor) ConvertOpenAIRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
 	if request == nil {
 		return nil, errors.New("request is nil")
 	}
@@ -60,6 +72,11 @@ func (a *Adaptor) ConvertEmbeddingRequest(c *gin.Context, info *relaycommon.Rela
 	return nil, errors.New("not implemented")
 }
 
+func (a *Adaptor) ConvertOpenAIResponsesRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.OpenAIResponsesRequest) (any, error) {
+	// TODO implement me
+	return nil, errors.New("not implemented")
+}
+
 func (a *Adaptor) DoRequest(c *gin.Context, info *relaycommon.RelayInfo, requestBody io.Reader) (any, error) {
 	return channel.DoApiRequest(a, c, info, requestBody)
 }
@@ -68,7 +85,7 @@ func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, info *relaycom
 	if info.IsStream {
 		err, usage = openai.OaiStreamHandler(c, resp, info)
 	} else {
-		err, usage = openai.OpenaiHandler(c, resp, info.PromptTokens, info.UpstreamModelName)
+		err, usage = openai.OpenaiHandler(c, resp, info)
 	}
 	return
 }

relay/channel/dify/adaptor.go

@@ -3,13 +3,13 @@ package dify
 import (
 	"errors"
 	"fmt"
-	"github.com/gin-gonic/gin"
 	"io"
 	"net/http"
 	"one-api/dto"
 	"one-api/relay/channel"
 	relaycommon "one-api/relay/common"
-	"strings"
+
+	"github.com/gin-gonic/gin"
 )
 
 const (
@@ -23,6 +23,12 @@ type Adaptor struct {
 	BotType int
 }
 
+func (a *Adaptor) ConvertClaudeRequest(*gin.Context, *relaycommon.RelayInfo, *dto.ClaudeRequest) (any, error) {
+	//TODO implement me
+	panic("implement me")
+	return nil, nil
+}
+
 func (a *Adaptor) ConvertAudioRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.AudioRequest) (io.Reader, error) {
 	//TODO implement me
 	return nil, errors.New("not implemented")
@@ -34,15 +40,16 @@ func (a *Adaptor) ConvertImageRequest(c *gin.Context, info *relaycommon.RelayInf
 }
 
 func (a *Adaptor) Init(info *relaycommon.RelayInfo) {
-	if strings.HasPrefix(info.UpstreamModelName, "agent") {
-		a.BotType = BotTypeAgent
-	} else if strings.HasPrefix(info.UpstreamModelName, "workflow") {
-		a.BotType = BotTypeWorkFlow
-	} else if strings.HasPrefix(info.UpstreamModelName, "chat") {
-		a.BotType = BotTypeCompletion
-	} else {
-		a.BotType = BotTypeChatFlow
-	}
+	//if strings.HasPrefix(info.UpstreamModelName, "agent") {
+	//	a.BotType = BotTypeAgent
+	//} else if strings.HasPrefix(info.UpstreamModelName, "workflow") {
+	//	a.BotType = BotTypeWorkFlow
+	//} else if strings.HasPrefix(info.UpstreamModelName, "chat") {
+	//	a.BotType = BotTypeCompletion
+	//} else {
+	//}
+	a.BotType = BotTypeChatFlow
+
 }
 
 func (a *Adaptor) GetRequestURL(info *relaycommon.RelayInfo) (string, error) {
@@ -64,11 +71,11 @@ func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *rel
 	return nil
 }
 
-func (a *Adaptor) ConvertRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
+func (a *Adaptor) ConvertOpenAIRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
 	if request == nil {
 		return nil, errors.New("request is nil")
 	}
-	return requestOpenAI2Dify(*request), nil
+	return requestOpenAI2Dify(c, info, *request), nil
 }
 
 func (a *Adaptor) ConvertRerankRequest(c *gin.Context, relayMode int, request dto.RerankRequest) (any, error) {
@@ -80,6 +87,11 @@ func (a *Adaptor) ConvertEmbeddingRequest(c *gin.Context, info *relaycommon.Rela
 	return nil, errors.New("not implemented")
 }
 
+func (a *Adaptor) ConvertOpenAIResponsesRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.OpenAIResponsesRequest) (any, error) {
+	// TODO implement me
+	return nil, errors.New("not implemented")
+}
+
 func (a *Adaptor) DoRequest(c *gin.Context, info *relaycommon.RelayInfo, requestBody io.Reader) (any, error) {
 	return channel.DoApiRequest(a, c, info, requestBody)
 }

relay/channel/dify/dto.go

@@ -8,6 +8,14 @@ type DifyChatRequest struct {
 	ResponseMode     string                 `json:"response_mode"`
 	User             string                 `json:"user"`
 	AutoGenerateName bool                   `json:"auto_generate_name"`
+	Files            []DifyFile             `json:"files"`
+}
+
+type DifyFile struct {
+	Type         string `json:"type"`
+	TransferMode string `json:"transfer_mode"`
+	URL          string `json:"url,omitempty"`
+	UploadFileId string `json:"upload_file_id,omitempty"`
 }
 
 type DifyMetaData struct {
@@ -17,6 +25,8 @@ type DifyMetaData struct {
 type DifyData struct {
 	WorkflowId string `json:"workflow_id"`
 	NodeId     string `json:"node_id"`
+	NodeType   string `json:"node_type"`
+	Status     string `json:"status"`
 }
 
 type DifyChatCompletionResponse struct {

relay/channel/dify/relay-dify.go

@@ -1,10 +1,12 @@
 package dify
 
 import (
-	"bufio"
+	"bytes"
+	"encoding/base64"
 	"encoding/json"
-	"github.com/gin-gonic/gin"
+	"fmt"
 	"io"
+	"mime/multipart"
 	"net/http"
 	"one-api/common"
 	"one-api/constant"
@@ -12,35 +14,163 @@ import (
 	relaycommon "one-api/relay/common"
 	"one-api/relay/helper"
 	"one-api/service"
+	"os"
 	"strings"
+
+	"github.com/gin-gonic/gin"
 )
 
-func requestOpenAI2Dify(request dto.GeneralOpenAIRequest) *DifyChatRequest {
-	content := ""
-	for _, message := range request.Messages {
-		if message.Role == "system" {
-			content += "SYSTEM: \n" + message.StringContent() + "\n"
-		} else if message.Role == "assistant" {
-			content += "ASSISTANT: \n" + message.StringContent() + "\n"
-		} else {
-			content += "USER: \n" + message.StringContent() + "\n"
+func uploadDifyFile(c *gin.Context, info *relaycommon.RelayInfo, user string, media dto.MediaContent) *DifyFile {
+	uploadUrl := fmt.Sprintf("%s/v1/files/upload", info.BaseUrl)
+	switch media.Type {
+	case dto.ContentTypeImageURL:
+		// Decode base64 data
+		imageMedia := media.GetImageMedia()
+		base64Data := imageMedia.Url
+		// Remove base64 prefix if exists (e.g., "data:image/jpeg;base64,")
+		if idx := strings.Index(base64Data, ","); idx != -1 {
+			base64Data = base64Data[idx+1:]
+		}
+
+		// Decode base64 string
+		decodedData, err := base64.StdEncoding.DecodeString(base64Data)
+		if err != nil {
+			common.SysError("failed to decode base64: " + err.Error())
+			return nil
+		}
+
+		// Create temporary file
+		tempFile, err := os.CreateTemp("", "dify-upload-*")
+		if err != nil {
+			common.SysError("failed to create temp file: " + err.Error())
+			return nil
+		}
+		defer tempFile.Close()
+		defer os.Remove(tempFile.Name())
+
+		// Write decoded data to temp file
+		if _, err := tempFile.Write(decodedData); err != nil {
+			common.SysError("failed to write to temp file: " + err.Error())
+			return nil
+		}
+
+		// Create multipart form
+		body := &bytes.Buffer{}
+		writer := multipart.NewWriter(body)
+
+		// Add user field
+		if err := writer.WriteField("user", user); err != nil {
+			common.SysError("failed to add user field: " + err.Error())
+			return nil
+		}
+
+		// Create form file with proper mime type
+		mimeType := imageMedia.MimeType
+		if mimeType == "" {
+			mimeType = "image/jpeg" // default mime type
+		}
+
+		// Create form file
+		part, err := writer.CreateFormFile("file", fmt.Sprintf("image.%s", strings.TrimPrefix(mimeType, "image/")))
+		if err != nil {
+			common.SysError("failed to create form file: " + err.Error())
+			return nil
+		}
+
+		// Copy file content to form
+		if _, err = io.Copy(part, bytes.NewReader(decodedData)); err != nil {
+			common.SysError("failed to copy file content: " + err.Error())
+			return nil
+		}
+		writer.Close()
+
+		// Create HTTP request
+		req, err := http.NewRequest("POST", uploadUrl, body)
+		if err != nil {
+			common.SysError("failed to create request: " + err.Error())
+			return nil
+		}
+
+		req.Header.Set("Content-Type", writer.FormDataContentType())
+		req.Header.Set("Authorization", fmt.Sprintf("Bearer %s", info.ApiKey))
+
+		// Send request
+		client := service.GetImpatientHttpClient()
+		resp, err := client.Do(req)
+		if err != nil {
+			common.SysError("failed to send request: " + err.Error())
+			return nil
+		}
+		defer resp.Body.Close()
+
+		// Parse response
+		var result struct {
+			Id string `json:"id"`
+		}
+		if err := json.NewDecoder(resp.Body).Decode(&result); err != nil {
+			common.SysError("failed to decode response: " + err.Error())
+			return nil
+		}
+
+		return &DifyFile{
+			UploadFileId: result.Id,
+			Type:         "image",
+			TransferMode: "local_file",
 		}
 	}
+	return nil
+}
+
+func requestOpenAI2Dify(c *gin.Context, info *relaycommon.RelayInfo, request dto.GeneralOpenAIRequest) *DifyChatRequest {
+	difyReq := DifyChatRequest{
+		Inputs:           make(map[string]interface{}),
+		AutoGenerateName: false,
+	}
+
+	user := request.User
+	if user == "" {
+		user = helper.GetResponseID(c)
+	}
+	difyReq.User = user
+
+	files := make([]DifyFile, 0)
+	var content strings.Builder
+	for _, message := range request.Messages {
+		if message.Role == "system" {
+			content.WriteString("SYSTEM: \n" + message.StringContent() + "\n")
+		} else if message.Role == "assistant" {
+			content.WriteString("ASSISTANT: \n" + message.StringContent() + "\n")
+		} else {
+			parseContent := message.ParseContent()
+			for _, mediaContent := range parseContent {
+				switch mediaContent.Type {
+				case dto.ContentTypeText:
+					content.WriteString("USER: \n" + mediaContent.Text + "\n")
+				case dto.ContentTypeImageURL:
+					media := mediaContent.GetImageMedia()
+					var file *DifyFile
+					if media.IsRemoteImage() {
+						file.Type = media.MimeType
+						file.TransferMode = "remote_url"
+						file.URL = media.Url
+					} else {
+						file = uploadDifyFile(c, info, difyReq.User, mediaContent)
+					}
+					if file != nil {
+						files = append(files, *file)
+					}
+				}
+			}
+		}
+	}
+	difyReq.Query = content.String()
+	difyReq.Files = files
 	mode := "blocking"
 	if request.Stream {
 		mode = "streaming"
 	}
-	user := request.User
-	if user == "" {
-		user = "api-user"
-	}
-	return &DifyChatRequest{
-		Inputs:           make(map[string]interface{}),
-		Query:            content,
-		ResponseMode:     mode,
-		User:             user,
-		AutoGenerateName: false,
-	}
+	difyReq.ResponseMode = mode
+	return &difyReq
 }
 
 func streamResponseDify2OpenAI(difyResponse DifyChunkChatCompletionResponse) *dto.ChatCompletionsStreamResponse {
@@ -50,11 +180,29 @@ func streamResponseDify2OpenAI(difyResponse DifyChunkChatCompletionResponse) *dt
 		Model:   "dify",
 	}
 	var choice dto.ChatCompletionsStreamResponseChoice
-	if constant.DifyDebug && difyResponse.Event == "workflow_started" {
-		choice.Delta.SetContentString("Workflow: " + difyResponse.Data.WorkflowId + "\n")
-	} else if constant.DifyDebug && difyResponse.Event == "node_started" {
-		choice.Delta.SetContentString("Node: " + difyResponse.Data.NodeId + "\n")
+	if strings.HasPrefix(difyResponse.Event, "workflow_") {
+		if constant.DifyDebug {
+			text := "Workflow: " + difyResponse.Data.WorkflowId
+			if difyResponse.Event == "workflow_finished" {
+				text += " " + difyResponse.Data.Status
+			}
+			choice.Delta.SetReasoningContent(text + "\n")
+		}
+	} else if strings.HasPrefix(difyResponse.Event, "node_") {
+		if constant.DifyDebug {
+			text := "Node: " + difyResponse.Data.NodeType
+			if difyResponse.Event == "node_finished" {
+				text += " " + difyResponse.Data.Status
+			}
+			choice.Delta.SetReasoningContent(text + "\n")
+		}
 	} else if difyResponse.Event == "message" || difyResponse.Event == "agent_message" {
+		if difyResponse.Answer == "<details style=\"color:gray;background-color: #f8f8f8;padding: 8px;border-radius: 4px;\" open> <summary> Thinking... </summary>\n" {
+			difyResponse.Answer = "<think>"
+		} else if difyResponse.Answer == "</details>" {
+			difyResponse.Answer = "</think>"
+		}
+
 		choice.Delta.SetContentString(difyResponse.Answer)
 	}
 	response.Choices = append(response.Choices, choice)
@@ -64,47 +212,40 @@ func streamResponseDify2OpenAI(difyResponse DifyChunkChatCompletionResponse) *dt
 func difyStreamHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (*dto.OpenAIErrorWithStatusCode, *dto.Usage) {
 	var responseText string
 	usage := &dto.Usage{}
-	scanner := bufio.NewScanner(resp.Body)
-	scanner.Split(bufio.ScanLines)
-
+	var nodeToken int
 	helper.SetEventStreamHeaders(c)
-
-	for scanner.Scan() {
-		data := scanner.Text()
-		if len(data) < 5 || !strings.HasPrefix(data, "data:") {
-			continue
-		}
-		data = strings.TrimPrefix(data, "data:")
+	helper.StreamScannerHandler(c, resp, info, func(data string) bool {
 		var difyResponse DifyChunkChatCompletionResponse
 		err := json.Unmarshal([]byte(data), &difyResponse)
 		if err != nil {
 			common.SysError("error unmarshalling stream response: " + err.Error())
-			continue
+			return true
 		}
 		var openaiResponse dto.ChatCompletionsStreamResponse
 		if difyResponse.Event == "message_end" {
 			usage = &difyResponse.MetaData.Usage
-			break
+			return false
 		} else if difyResponse.Event == "error" {
-			break
+			return false
 		} else {
 			openaiResponse = *streamResponseDify2OpenAI(difyResponse)
 			if len(openaiResponse.Choices) != 0 {
 				responseText += openaiResponse.Choices[0].Delta.GetContentString()
+				if openaiResponse.Choices[0].Delta.ReasoningContent != nil {
+					nodeToken += 1
+				}
 			}
 		}
 		err = helper.ObjectData(c, openaiResponse)
 		if err != nil {
 			common.SysError(err.Error())
 		}
-	}
-	if err := scanner.Err(); err != nil {
-		common.SysError("error reading stream: " + err.Error())
-	}
+		return true
+	})
 	helper.Done(c)
 	err := resp.Body.Close()
 	if err != nil {
-		//return service.OpenAIErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
+		// return service.OpenAIErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError), nil
 		common.SysError("close_response_body_failed: " + err.Error())
 	}
 	if usage.TotalTokens == 0 {
@@ -112,6 +253,7 @@ func difyStreamHandler(c *gin.Context, resp *http.Response, info *relaycommon.Re
 		usage.CompletionTokens, _ = service.CountTextToken("gpt-3.5-turbo", responseText)
 		usage.TotalTokens = usage.PromptTokens + usage.CompletionTokens
 	}
+	usage.CompletionTokens += nodeToken
 	return nil, usage
 }
 
@@ -136,12 +278,11 @@ func difyHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInf
 		Created: common.GetTimestamp(),
 		Usage:   difyResponse.MetaData.Usage,
 	}
-	content, _ := json.Marshal(difyResponse.Answer)
 	choice := dto.OpenAITextResponseChoice{
 		Index: 0,
 		Message: dto.Message{
 			Role:    "assistant",
-			Content: content,
+			Content: difyResponse.Answer,
 		},
 		FinishReason: "stop",
 	}

relay/channel/gemini/adaptor.go

@@ -10,9 +10,9 @@ import (
 	"one-api/dto"
 	"one-api/relay/channel"
 	relaycommon "one-api/relay/common"
+	"one-api/relay/constant"
 	"one-api/service"
 	"one-api/setting/model_setting"
-
 	"strings"
 
 	"github.com/gin-gonic/gin"
@@ -21,6 +21,12 @@ import (
 type Adaptor struct {
 }
 
+func (a *Adaptor) ConvertClaudeRequest(*gin.Context, *relaycommon.RelayInfo, *dto.ClaudeRequest) (any, error) {
+	//TODO implement me
+	panic("implement me")
+	return nil, nil
+}
+
 func (a *Adaptor) ConvertAudioRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.AudioRequest) (io.Reader, error) {
 	//TODO implement me
 	return nil, errors.New("not implemented")
@@ -64,12 +70,28 @@ func (a *Adaptor) Init(info *relaycommon.RelayInfo) {
 }
 
 func (a *Adaptor) GetRequestURL(info *relaycommon.RelayInfo) (string, error) {
+
+	if model_setting.GetGeminiSettings().ThinkingAdapterEnabled {
+		// suffix -thinking and -nothinking
+		if strings.HasSuffix(info.OriginModelName, "-thinking") {
+			info.UpstreamModelName = strings.TrimSuffix(info.UpstreamModelName, "-thinking")
+		} else if strings.HasSuffix(info.OriginModelName, "-nothinking") {
+			info.UpstreamModelName = strings.TrimSuffix(info.UpstreamModelName, "-nothinking")
+		}
+	}
+
 	version := model_setting.GetGeminiVersionSetting(info.UpstreamModelName)
 
 	if strings.HasPrefix(info.UpstreamModelName, "imagen") {
 		return fmt.Sprintf("%s/%s/models/%s:predict", info.BaseUrl, version, info.UpstreamModelName), nil
 	}
 
+	if strings.HasPrefix(info.UpstreamModelName, "text-embedding") ||
+		strings.HasPrefix(info.UpstreamModelName, "embedding") ||
+		strings.HasPrefix(info.UpstreamModelName, "gemini-embedding") {
+		return fmt.Sprintf("%s/%s/models/%s:embedContent", info.BaseUrl, version, info.UpstreamModelName), nil
+	}
+
 	action := "generateContent"
 	if info.IsStream {
 		action = "streamGenerateContent?alt=sse"
@@ -83,15 +105,17 @@ func (a *Adaptor) SetupRequestHeader(c *gin.Context, req *http.Header, info *rel
 	return nil
 }
 
-func (a *Adaptor) ConvertRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
+func (a *Adaptor) ConvertOpenAIRequest(c *gin.Context, info *relaycommon.RelayInfo, request *dto.GeneralOpenAIRequest) (any, error) {
 	if request == nil {
 		return nil, errors.New("request is nil")
 	}
-	ai, err := CovertGemini2OpenAI(*request)
+
+	geminiRequest, err := CovertGemini2OpenAI(*request, info)
 	if err != nil {
 		return nil, err
 	}
-	return ai, nil
+
+	return geminiRequest, nil
 }
 
 func (a *Adaptor) ConvertRerankRequest(c *gin.Context, relayMode int, request dto.RerankRequest) (any, error) {
@@ -99,7 +123,41 @@ func (a *Adaptor) ConvertRerankRequest(c *gin.Context, relayMode int, request dt
 }
 
 func (a *Adaptor) ConvertEmbeddingRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.EmbeddingRequest) (any, error) {
-	//TODO implement me
+	if request.Input == nil {
+		return nil, errors.New("input is required")
+	}
+
+	inputs := request.ParseInput()
+	if len(inputs) == 0 {
+		return nil, errors.New("input is empty")
+	}
+
+	// only process the first input
+	geminiRequest := GeminiEmbeddingRequest{
+		Content: GeminiChatContent{
+			Parts: []GeminiPart{
+				{
+					Text: inputs[0],
+				},
+			},
+		},
+	}
+
+	// set specific parameters for different models
+	// https://ai.google.dev/api/embeddings?hl=zh-cn#method:-models.embedcontent
+	switch info.UpstreamModelName {
+	case "text-embedding-004":
+		// except embedding-001 supports setting `OutputDimensionality`
+		if request.Dimensions > 0 {
+			geminiRequest.OutputDimensionality = request.Dimensions
+		}
+	}
+
+	return geminiRequest, nil
+}
+
+func (a *Adaptor) ConvertOpenAIResponsesRequest(c *gin.Context, info *relaycommon.RelayInfo, request dto.OpenAIResponsesRequest) (any, error) {
+	// TODO implement me
 	return nil, errors.New("not implemented")
 }
 
@@ -108,15 +166,42 @@ func (a *Adaptor) DoRequest(c *gin.Context, info *relaycommon.RelayInfo, request
 }
 
 func (a *Adaptor) DoResponse(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (usage any, err *dto.OpenAIErrorWithStatusCode) {
+	if info.RelayMode == constant.RelayModeGemini {
+		if info.IsStream {
+			return GeminiTextGenerationStreamHandler(c, resp, info)
+		} else {
+			return GeminiTextGenerationHandler(c, resp, info)
+		}
+	}
+
 	if strings.HasPrefix(info.UpstreamModelName, "imagen") {
 		return GeminiImageHandler(c, resp, info)
 	}
 
+	// check if the model is an embedding model
+	if strings.HasPrefix(info.UpstreamModelName, "text-embedding") ||
+		strings.HasPrefix(info.UpstreamModelName, "embedding") ||
+		strings.HasPrefix(info.UpstreamModelName, "gemini-embedding") {
+		return GeminiEmbeddingHandler(c, resp, info)
+	}
+
 	if info.IsStream {
 		err, usage = GeminiChatStreamHandler(c, resp, info)
 	} else {
 		err, usage = GeminiChatHandler(c, resp, info)
 	}
+
+	//if usage.(*dto.Usage).CompletionTokenDetails.ReasoningTokens > 100 {
+	//	// 没有请求-thinking的情况下，产生思考token，则按照思考模型计费
+	//	if !strings.HasSuffix(info.OriginModelName, "-thinking") &&
+	//		!strings.HasSuffix(info.OriginModelName, "-nothinking") {
+	//		thinkingModelName := info.OriginModelName + "-thinking"
+	//		if operation_setting.SelfUseModeEnabled || helper.ContainPriceOrRatio(thinkingModelName) {
+	//			info.OriginModelName = thinkingModelName
+	//		}
+	//	}
+	//}
+
 	return
 }
 

relay/channel/gemini/constant.go

@@ -16,8 +16,14 @@ var ModelList = []string{
 	"gemini-2.0-pro-exp",
 	// thinking exp
 	"gemini-2.0-flash-thinking-exp",
+	"gemini-2.5-pro-exp-03-25",
+	"gemini-2.5-pro-preview-03-25",
 	// imagen models
 	"imagen-3.0-generate-002",
+	// embedding models
+	"gemini-embedding-exp-03-07",
+	"text-embedding-004",
+	"embedding-001",
 }
 
 var SafetySettingList = []string{

relay/channel/gemini/dto.go

@@ -1,11 +1,22 @@
 package gemini
 
+import "encoding/json"
+
 type GeminiChatRequest struct {
 	Contents           []GeminiChatContent        `json:"contents"`
-	SafetySettings     []GeminiChatSafetySettings `json:"safety_settings,omitempty"`
-	GenerationConfig   GeminiChatGenerationConfig `json:"generation_config,omitempty"`
+	SafetySettings     []GeminiChatSafetySettings `json:"safetySettings,omitempty"`
+	GenerationConfig   GeminiChatGenerationConfig `json:"generationConfig,omitempty"`
 	Tools              []GeminiChatTool           `json:"tools,omitempty"`
-	SystemInstructions *GeminiChatContent         `json:"system_instruction,omitempty"`
+	SystemInstructions *GeminiChatContent         `json:"systemInstruction,omitempty"`
+}
+
+type GeminiThinkingConfig struct {
+	IncludeThoughts bool `json:"includeThoughts,omitempty"`
+	ThinkingBudget  *int `json:"thinkingBudget,omitempty"`
+}
+
+func (c *GeminiThinkingConfig) SetThinkingBudget(budget int) {
+	c.ThinkingBudget = &budget
 }
 
 type GeminiInlineData struct {
@@ -13,19 +24,38 @@ type GeminiInlineData struct {
 	Data     string `json:"data"`
 }
 
+// UnmarshalJSON custom unmarshaler for GeminiInlineData to support snake_case and camelCase for MimeType
+func (g *GeminiInlineData) UnmarshalJSON(data []byte) error {
+	type Alias GeminiInlineData // Use type alias to avoid recursion
+	var aux struct {
+		Alias
+		MimeTypeSnake string `json:"mime_type"`
+	}
+
+	if err := json.Unmarshal(data, &aux); err != nil {
+		return err
+	}
+
+	*g = GeminiInlineData(aux.Alias) // Copy other fields if any in future
+
+	// Prioritize snake_case if present
+	if aux.MimeTypeSnake != "" {
+		g.MimeType = aux.MimeTypeSnake
+	} else if aux.MimeType != "" { // Fallback to camelCase from Alias
+		g.MimeType = aux.MimeType
+	}
+	// g.Data would be populated by aux.Alias.Data
+	return nil
+}
+
 type FunctionCall struct {
 	FunctionName string `json:"name"`
 	Arguments    any    `json:"args"`
 }
 
-type GeminiFunctionResponseContent struct {
-	Name    string `json:"name"`
-	Content any    `json:"content"`
-}
-
 type FunctionResponse struct {
-	Name     string                        `json:"name"`
-	Response GeminiFunctionResponseContent `json:"response"`
+	Name     string                 `json:"name"`
+	Response map[string]interface{} `json:"response"`
 }
 
 type GeminiPartExecutableCode struct {
@@ -45,6 +75,7 @@ type GeminiFileData struct {
 
 type GeminiPart struct {
 	Text                string                         `json:"text,omitempty"`
+	Thought             bool                           `json:"thought,omitempty"`
 	InlineData          *GeminiInlineData              `json:"inlineData,omitempty"`
 	FunctionCall        *FunctionCall                  `json:"functionCall,omitempty"`
 	FunctionResponse    *FunctionResponse              `json:"functionResponse,omitempty"`
@@ -53,6 +84,33 @@ type GeminiPart struct {
 	CodeExecutionResult *GeminiPartCodeExecutionResult `json:"codeExecutionResult,omitempty"`
 }
 
+// UnmarshalJSON custom unmarshaler for GeminiPart to support snake_case and camelCase for InlineData
+func (p *GeminiPart) UnmarshalJSON(data []byte) error {
+	// Alias to avoid recursion during unmarshalling
+	type Alias GeminiPart
+	var aux struct {
+		Alias
+		InlineDataSnake *GeminiInlineData `json:"inline_data,omitempty"` // snake_case variant
+	}
+
+	if err := json.Unmarshal(data, &aux); err != nil {
+		return err
+	}
+
+	// Assign fields from alias
+	*p = GeminiPart(aux.Alias)
+
+	// Prioritize snake_case for InlineData if present
+	if aux.InlineDataSnake != nil {
+		p.InlineData = aux.InlineDataSnake
+	} else if aux.InlineData != nil { // Fallback to camelCase from Alias
+		p.InlineData = aux.InlineData
+	}
+	// Other fields like Text, FunctionCall etc. are already populated via aux.Alias
+
+	return nil
+}
+
 type GeminiChatContent struct {
 	Role  string       `json:"role,omitempty"`
 	Parts []GeminiPart `json:"parts"`
@@ -71,15 +129,17 @@ type GeminiChatTool struct {
 }
 
 type GeminiChatGenerationConfig struct {
-	Temperature      *float64 `json:"temperature,omitempty"`
-	TopP             float64  `json:"topP,omitempty"`
-	TopK             float64  `json:"topK,omitempty"`
-	MaxOutputTokens  uint     `json:"maxOutputTokens,omitempty"`
-	CandidateCount   int      `json:"candidateCount,omitempty"`
-	StopSequences    []string `json:"stopSequences,omitempty"`
-	ResponseMimeType string   `json:"responseMimeType,omitempty"`
-	ResponseSchema   any      `json:"responseSchema,omitempty"`
-	Seed             int64    `json:"seed,omitempty"`
+	Temperature        *float64              `json:"temperature,omitempty"`
+	TopP               float64               `json:"topP,omitempty"`
+	TopK               float64               `json:"topK,omitempty"`
+	MaxOutputTokens    uint                  `json:"maxOutputTokens,omitempty"`
+	CandidateCount     int                   `json:"candidateCount,omitempty"`
+	StopSequences      []string              `json:"stopSequences,omitempty"`
+	ResponseMimeType   string                `json:"responseMimeType,omitempty"`
+	ResponseSchema     any                   `json:"responseSchema,omitempty"`
+	Seed               int64                 `json:"seed,omitempty"`
+	ResponseModalities []string              `json:"responseModalities,omitempty"`
+	ThinkingConfig     *GeminiThinkingConfig `json:"thinkingConfig,omitempty"`
 }
 
 type GeminiChatCandidate struct {
@@ -105,9 +165,16 @@ type GeminiChatResponse struct {
 }
 
 type GeminiUsageMetadata struct {
-	PromptTokenCount     int `json:"promptTokenCount"`
-	CandidatesTokenCount int `json:"candidatesTokenCount"`
-	TotalTokenCount      int `json:"totalTokenCount"`
+	PromptTokenCount     int                         `json:"promptTokenCount"`
+	CandidatesTokenCount int                         `json:"candidatesTokenCount"`
+	TotalTokenCount      int                         `json:"totalTokenCount"`
+	ThoughtsTokenCount   int                         `json:"thoughtsTokenCount"`
+	PromptTokensDetails  []GeminiPromptTokensDetails `json:"promptTokensDetails"`
+}
+
+type GeminiPromptTokensDetails struct {
+	Modality   string `json:"modality"`
+	TokenCount int    `json:"tokenCount"`
 }
 
 // Imagen related structs
@@ -136,3 +203,19 @@ type GeminiImagePrediction struct {
 	RaiFilteredReason  string `json:"raiFilteredReason,omitempty"`
 	SafetyAttributes   any    `json:"safetyAttributes,omitempty"`
 }
+
+// Embedding related structs
+type GeminiEmbeddingRequest struct {
+	Content              GeminiChatContent `json:"content"`
+	TaskType             string            `json:"taskType,omitempty"`
+	Title                string            `json:"title,omitempty"`
+	OutputDimensionality int               `json:"outputDimensionality,omitempty"`
+}
+
+type GeminiEmbeddingResponse struct {
+	Embedding ContentEmbedding `json:"embedding"`
+}
+
+type ContentEmbedding struct {
+	Values []float64 `json:"values"`
+}

relay/channel/gemini/relay-gemini-native.go

@@ -0,0 +1,145 @@
+package gemini
+
+import (
+	"encoding/json"
+	"io"
+	"net/http"
+	"one-api/common"
+	"one-api/dto"
+	relaycommon "one-api/relay/common"
+	"one-api/relay/helper"
+	"one-api/service"
+
+	"github.com/gin-gonic/gin"
+)
+
+func GeminiTextGenerationHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (*dto.Usage, *dto.OpenAIErrorWithStatusCode) {
+	// 读取响应体
+	responseBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return nil, service.OpenAIErrorWrapper(err, "read_response_body_failed", http.StatusInternalServerError)
+	}
+	err = resp.Body.Close()
+	if err != nil {
+		return nil, service.OpenAIErrorWrapper(err, "close_response_body_failed", http.StatusInternalServerError)
+	}
+
+	if common.DebugEnabled {
+		println(string(responseBody))
+	}
+
+	// 解析为 Gemini 原生响应格式
+	var geminiResponse GeminiChatResponse
+	err = common.DecodeJson(responseBody, &geminiResponse)
+	if err != nil {
+		return nil, service.OpenAIErrorWrapper(err, "unmarshal_response_body_failed", http.StatusInternalServerError)
+	}
+
+	// 检查是否有候选响应
+	if len(geminiResponse.Candidates) == 0 {
+		return nil, &dto.OpenAIErrorWithStatusCode{
+			Error: dto.OpenAIError{
+				Message: "No candidates returned",
+				Type:    "server_error",
+				Param:   "",
+				Code:    500,
+			},
+			StatusCode: resp.StatusCode,
+		}
+	}
+
+	// 计算使用量（基于 UsageMetadata）
+	usage := dto.Usage{
+		PromptTokens:     geminiResponse.UsageMetadata.PromptTokenCount,
+		CompletionTokens: geminiResponse.UsageMetadata.CandidatesTokenCount,
+		TotalTokens:      geminiResponse.UsageMetadata.TotalTokenCount,
+	}
+
+	usage.CompletionTokenDetails.ReasoningTokens = geminiResponse.UsageMetadata.ThoughtsTokenCount
+
+	for _, detail := range geminiResponse.UsageMetadata.PromptTokensDetails {
+		if detail.Modality == "AUDIO" {
+			usage.PromptTokensDetails.AudioTokens = detail.TokenCount
+		} else if detail.Modality == "TEXT" {
+			usage.PromptTokensDetails.TextTokens = detail.TokenCount
+		}
+	}
+
+	// 直接返回 Gemini 原生格式的 JSON 响应
+	jsonResponse, err := json.Marshal(geminiResponse)
+	if err != nil {
+		return nil, service.OpenAIErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError)
+	}
+
+	// 设置响应头并写入响应
+	c.Writer.Header().Set("Content-Type", "application/json")
+	c.Writer.WriteHeader(resp.StatusCode)
+	_, err = c.Writer.Write(jsonResponse)
+	if err != nil {
+		return nil, service.OpenAIErrorWrapper(err, "write_response_failed", http.StatusInternalServerError)
+	}
+
+	return &usage, nil
+}
+
+func GeminiTextGenerationStreamHandler(c *gin.Context, resp *http.Response, info *relaycommon.RelayInfo) (*dto.Usage, *dto.OpenAIErrorWithStatusCode) {
+	var usage = &dto.Usage{}
+	var imageCount int
+
+	helper.SetEventStreamHeaders(c)
+
+	helper.StreamScannerHandler(c, resp, info, func(data string) bool {
+		var geminiResponse GeminiChatResponse
+		err := common.DecodeJsonStr(data, &geminiResponse)
+		if err != nil {
+			common.LogError(c, "error unmarshalling stream response: "+err.Error())
+			return false
+		}
+
+		// 统计图片数量
+		for _, candidate := range geminiResponse.Candidates {
+			for _, part := range candidate.Content.Parts {
+				if part.InlineData != nil && part.InlineData.MimeType != "" {
+					imageCount++
+				}
+			}
+		}
+
+		// 更新使用量统计
+		if geminiResponse.UsageMetadata.TotalTokenCount != 0 {
+			usage.PromptTokens = geminiResponse.UsageMetadata.PromptTokenCount
+			usage.CompletionTokens = geminiResponse.UsageMetadata.CandidatesTokenCount
+			usage.TotalTokens = geminiResponse.UsageMetadata.TotalTokenCount
+			usage.CompletionTokenDetails.ReasoningTokens = geminiResponse.UsageMetadata.ThoughtsTokenCount
+			for _, detail := range geminiResponse.UsageMetadata.PromptTokensDetails {
+				if detail.Modality == "AUDIO" {
+					usage.PromptTokensDetails.AudioTokens = detail.TokenCount
+				} else if detail.Modality == "TEXT" {
+					usage.PromptTokensDetails.TextTokens = detail.TokenCount
+				}
+			}
+		}
+
+		// 直接发送 GeminiChatResponse 响应
+		err = helper.ObjectData(c, geminiResponse)
+		if err != nil {
+			common.LogError(c, err.Error())
+		}
+
+		return true
+	})
+
+	if imageCount != 0 {
+		if usage.CompletionTokens == 0 {
+			usage.CompletionTokens = imageCount * 258
+		}
+	}
+
+	// 计算最终使用量
+	usage.CompletionTokens = usage.TotalTokens - usage.PromptTokens
+
+	// 移除流式响应结尾的[Done]，因为Gemini API没有发送Done的行为
+	//helper.Done(c)
+
+	return usage, nil
+}