多渠道AI API统一转换代理系统

中文版

📖 项目概述

这是一个多渠道AI API统一转换代理系统，支持OpenAI、Anthropic Claude、Google Gemini三种API格式的相互转换，具备多渠道管理和全面能力检测功能。

🔄 系统工作原理

核心转换流程

graph LR
    A[客户端请求] --> B{格式识别}
    B --> C[渠道路由]
    C --> D[格式转换]
    D --> E[请求转发]
    E --> F[AI服务API]
    F --> G[响应转换]
    G --> H[返回客户端]

🎯 1. 格式识别

自动检测：根据请求路径和参数自动识别源API格式
支持格式：OpenAI /v1/chat/completions | Anthropic /v1/messages | Gemini /v1/models
智能解析：解析请求头、参数结构，确定源格式规范

🚀 2. 渠道路由

Key映射：根据自定义API Key查找目标渠道配置
负载均衡：支持多渠道轮询和权重分配
故障转移：自动切换到备用渠道，确保服务可用性

⚡ 3. 格式转换

请求转换：将源格式的请求体转换为目标API格式
参数映射：自动处理模型名称、参数结构的差异
兼容处理：保持所有高级功能的完整性

🌐 4. 请求转发

HTTP代理：透明转发到真实的AI服务API
认证处理：自动注入目标渠道的API Key和认证信息
超时控制：可配置的请求超时和重试机制

🔄 5. 响应转换

格式统一：将目标API响应转换回源格式
流式支持：完整支持SSE流式响应的格式转换
错误映射：统一错误码和错误信息格式

🎯 核心功能

1. 全面能力检测

基础能力：聊天对话、流式输出、系统消息、多轮对话
高级能力：视觉理解、文件上传、结构化输出、JSON模式
工具能力：函数调用、工具使用、代码执行
模型检测：自动获取支持的模型列表
多平台支持：OpenAI、Anthropic Claude、Google Gemini

2. 智能格式转换

# 支持的转换路径
OpenAI ↔ Anthropic ↔ Gemini
  ↑         ↑         ↑
  └─────────┼─────────┘
            │
        任意互转

支持的高级功能转换：

✅ 流式响应：SSE格式的完整转换
✅ 函数调用：Tool Calling跨平台映射
✅ 视觉理解：图像输入格式统一处理
✅ 结构化输出：JSON Schema自动适配
✅ 模型映射：智能模型名称转换

3. 多渠道管理（规划中）

渠道类型：官方API、代理服务、自建服务
智能路由：基于延迟、成功率的智能选择
健康检查：实时监控渠道可用性
配额管理：支持多渠道的配额分配

🚀 快速开始

安装依赖

pip install -r requirements.txt

启动Web服务

python web_server.py

访问Web界面

打开浏览器访问：http://localhost:3000
选择AI提供商，输入API配置
一键检测所有能力，查看详细结果
使用转换功能，详见系统工作原理

🔧 客户端集成指南

Claude Code 中使用

Mac

export ANTHROPIC_BASE_URL="https://your_url.com"
# 测试发现claude code密钥需要以sk-开头
export ANTHROPIC_API_KEY="sk-xxx"
claude --model your_model

Windows CMD

set ANTHROPIC_BASE_URL=https://your_url.com
# 测试发现claude code密钥需要以sk-开头
set ANTHROPIC_API_KEY=sk-xxx
claude --model your_model

Gemini-CLI 中使用

Mac

export GOOGLE_GEMINI_BASE_URL="https://your_url.com"
export GEMINI_API_KEY="your_api_key"
gemini -m your_model

Windows CMD

set GOOGLE_GEMINI_BASE_URL=https://your_url.com
set GEMINI_API_KEY=your_api_key
gemini -m your_model

Cherry Studio 中使用

选择你想转换的供应商格式，填入url，填入你想使用的渠道的key

🚢 部署指南

Render 平台部署（推荐）

项目已配置好 render.yaml，支持一键部署：

将代码推送到GitHub
连接Render平台：https://dashboard.render.com
自动部署：Render会自动读取配置并部署

配置详情：

构建命令：pip install -r requirements.txt
启动命令：python web_server.py --host 0.0.0.0 --port $PORT
环境变量：PYTHONPATH=/opt/render/project/src

Docker 部署

# 构建镜像
docker build -t ai-api-detector .

# 运行容器  
docker run -p 8000:8000 ai-api-detector

本地开发

# 克隆项目
git clone <repository-url>
cd Api-Conversion

# 安装依赖
pip install -r requirements.txt

# 启动开发服务器
python web_server.py --debug

📊 支持的能力检测

能力	描述	OpenAI	Anthropic	Gemini
基础聊天	基本对话功能	✅	✅	✅
流式输出	实时流式响应	✅	✅	✅
系统消息	系统指令支持	✅	✅	✅
函数调用	工具使用能力	✅	✅	✅
结构化输出	JSON格式输出	✅	✅	✅
视觉理解	图像分析能力	✅	✅	✅

📄 许可证

MIT License

English Version

🇺🇸 English | 🇨🇳 中文

Multi-Channel AI API Unified Conversion Proxy System

📖 Project Overview

This is a multi-channel AI API unified conversion proxy system that supports mutual conversion between OpenAI, Anthropic Claude, and Google Gemini API formats, with multi-channel management and comprehensive capability detection features.

🔄 System Working Principle

Core Conversion Flow

graph LR
    A[Client Request] --> B{Format Recognition}
    B --> C[Channel Routing]
    C --> D[Format Conversion]
    D --> E[Request Forwarding]
    E --> F[AI Service API]
    F --> G[Response Conversion]
    G --> H[Return to Client]

🎯 1. Format Recognition

Auto Detection: Automatically identify source API format based on request path and parameters
Supported Formats: OpenAI /v1/chat/completions | Anthropic /v1/messages | Gemini /v1/models
Smart Parsing: Parse request headers and parameter structure to determine source format specification

🚀 2. Channel Routing

Key Mapping: Find target channel configuration based on custom API Key
Load Balancing: Support multi-channel polling and weight allocation
Failover: Automatically switch to backup channels to ensure service availability

⚡ 3. Format Conversion

Request Conversion: Convert source format request body to target API format
Parameter Mapping: Automatically handle differences in model names and parameter structures
Compatibility Handling: Maintain integrity of all advanced features

🌐 4. Request Forwarding

HTTP Proxy: Transparently forward to real AI service APIs
Authentication Handling: Automatically inject target channel's API Key and authentication info
Timeout Control: Configurable request timeout and retry mechanisms

🔄 5. Response Conversion

Format Unification: Convert target API response back to source format
Streaming Support: Full support for SSE streaming response format conversion
Error Mapping: Unified error codes and error message formats

🎯 Core Features

1. Comprehensive Capability Detection

Basic Capabilities: Chat conversation, streaming output, system messages, multi-turn dialogue
Advanced Capabilities: Vision understanding, file upload, structured output, JSON mode
Tool Capabilities: Function calling, tool usage, code execution
Model Detection: Automatically retrieve supported model lists
Multi-Platform Support: OpenAI, Anthropic Claude, Google Gemini

2. Smart Format Conversion

# Supported conversion paths
OpenAI ↔ Anthropic ↔ Gemini
  ↑         ↑         ↑
  └─────────┼─────────┘
            │
     Any mutual conversion

Supported advanced feature conversions:

✅ Streaming Response: Complete conversion of SSE format
✅ Function Calling: Tool Calling cross-platform mapping
✅ Vision Understanding: Unified processing of image input formats
✅ Structured Output: Automatic JSON Schema adaptation
✅ Model Mapping: Smart model name conversion

3. Multi-Channel Management (Planned)

Channel Types: Official API, proxy services, self-hosted services
Smart Routing: Intelligent selection based on latency and success rate
Health Checks: Real-time monitoring of channel availability
Quota Management: Support quota allocation across multiple channels

🚀 Quick Start

Install Dependencies

pip install -r requirements.txt

Start Web Service

python web_server.py

Access Web Interface

Open browser and visit: http://localhost:3000
Select AI provider, input API configuration
One-click detection of all capabilities, view detailed results
Use conversion functionality, see system working principle for details

🔧 Client Integration Guide

Using with Claude Code

Mac

export ANTHROPIC_BASE_URL="https://your_url.com"
# Testing found that claude code requires keys to start with sk-
export ANTHROPIC_API_KEY="sk-xxx"
claude --model your_model

Windows CMD

set ANTHROPIC_BASE_URL=https://your_url.com
# Testing found that claude code requires keys to start with sk-
set ANTHROPIC_API_KEY=sk-xxx
claude --model your_model

Using with Gemini-CLI

Mac

export GOOGLE_GEMINI_BASE_URL="https://your_url.com"
export GEMINI_API_KEY="your_api_key"
gemini -m your_model

Windows CMD

set GOOGLE_GEMINI_BASE_URL=https://your_url.com
set GEMINI_API_KEY=your_api_key
gemini -m your_model

Using with Cherry Studio

Select the provider format you want to convert, fill in the URL, and enter the key for the channel you want to use

🚢 Deployment Guide

Render Platform Deployment (Recommended)

The project is pre-configured with render.yaml for one-click deployment:

Push code to GitHub
Connect Render platform: https://dashboard.render.com
Auto Deploy: Render will automatically read configuration and deploy

Configuration Details:

Build Command: pip install -r requirements.txt
Start Command: python web_server.py --host 0.0.0.0 --port $PORT
Environment Variables: PYTHONPATH=/opt/render/project/src

Docker Deployment

# Build image
docker build -t ai-api-detector .

# Run container
docker run -p 8000:8000 ai-api-detector

Local Development

# Clone project
git clone <repository-url>
cd Api-Conversion

# Install dependencies
pip install -r requirements.txt

# Start development server
python web_server.py --debug

📊 Supported Capability Detection

Capability	Description	OpenAI	Anthropic	Gemini
Basic Chat	Basic conversation functionality	✅	✅	✅
Streaming Output	Real-time streaming response	✅	✅	✅
System Messages	System instruction support	✅	✅	✅
Function Calling	Tool usage capability	✅	✅	✅
Structured Output	JSON format output	✅	✅	✅
Vision Understanding	Image analysis capability	✅	✅	✅

📄 License

MIT License

Name		Name	Last commit message	Last commit date
Latest commit History 5 Commits
images		images
src		src
static		static
.dockerignore		.dockerignore
.env.example		.env.example
.gitignore		.gitignore
Dockerfile		Dockerfile
README.md		README.md
render.yaml		render.yaml
requirements.txt		requirements.txt
web_server.py		web_server.py

fishforks/Api-Conversion

Folders and files

Latest commit

History

Repository files navigation

多渠道AI API统一转换代理系统

中文版

📖 项目概述

核心转换流程

🎯 1. 格式识别

🚀 2. 渠道路由

⚡ 3. 格式转换

🌐 4. 请求转发

🔄 5. 响应转换

🎯 核心功能

1. 全面能力检测

2. 智能格式转换

3. 多渠道管理（规划中）

🚀 快速开始

🔧 客户端集成指南

Claude Code 中使用

Mac

Windows CMD

Gemini-CLI 中使用

Mac

Windows CMD

Cherry Studio 中使用

🚢 部署指南

Render 平台部署（推荐）

Docker 部署

本地开发

📊 支持的能力检测

📄 许可证

English Version

Multi-Channel AI API Unified Conversion Proxy System

📖 Project Overview

Core Conversion Flow

🎯 1. Format Recognition

🚀 2. Channel Routing

⚡ 3. Format Conversion

🌐 4. Request Forwarding

🔄 5. Response Conversion

🎯 Core Features

1. Comprehensive Capability Detection

2. Smart Format Conversion

3. Multi-Channel Management (Planned)

🚀 Quick Start

🔧 Client Integration Guide

Using with Claude Code

Mac

Windows CMD

Using with Gemini-CLI

Mac

Windows CMD

Using with Cherry Studio

🚢 Deployment Guide

Render Platform Deployment (Recommended)

Docker Deployment

Local Development

📊 Supported Capability Detection

📄 License

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Languages

Packages