endykaufman/qwen-api-proxy repository overview

⁠FreeQwenApi

Free Qwen AI API Proxy with OpenAI compatibility and Telegram bot integration. Access Qwen models without API keys.

⁠✨ Features

• 25+ Qwen Models: qwen3-max, qwen3.5-plus, qwen3-coder-plus, qwen-vl-plus, and more
• OpenAI Compatible: Works with OpenAI SDK, just change baseURL
• Telegram Bot: Session upload, LLM chat, health monitoring
• Streaming Support: Server-Sent Events (SSE) for real-time responses
• File Upload: Upload images and documents for analysis
• Image Generation: DALL-E compatible interface via Qwen Image API
• Multi-Account: Automatic rotation with rate limit handling
• Context Management: Server-side chat history with chatId/parentId

⁠🚀 Quick Start

⁠Docker (Recommended)

# 1. Add account locally first
npm run auth

# 2. Build and run
docker compose build --no-cache
docker compose up -d

⁠Docker Compose

services:
  qwen-proxy:
    build: .
    container_name: qwen-proxy
    environment:
      - SKIP_ACCOUNT_MENU=true
      - PORT=3264
    ports:
      - "3264:3264"
    volumes:
      - ./session:/app/session
      - ./logs:/app/logs
      - ./uploads:/app/uploads
    restart: unless-stopped

⁠Node.js

git clone https://github.com/EndyKaufman/FreeQwenApi
cd FreeQwenApi
npm install
npm start

⁠📡 API Endpoints

⁠💡 Usage Examples

⁠Basic Chat Request

curl http://localhost:3264/api/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model":"qwen-max-latest","messages":[{"role":"user","content":"Hello!"}]}'

⁠OpenAI SDK

import OpenAI from 'openai';

const client = new OpenAI({
    baseURL: 'http://localhost:3264/api',
    apiKey: 'any-string',
});

const response = await client.chat.completions.create({
    model: 'qwen-max-latest',
    messages: [{ role: 'user', content: 'Hello!' }],
});

⁠Streaming

curl -X POST http://localhost:3264/api/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model":"qwen-max-latest","messages":[{"role":"user","content":"Write a poem"}],"stream":true}'

⁠Context-Aware Conversation

// First message
const res1 = await fetch('/api/chat', {
  method: 'POST',
  body: JSON.stringify({ message: "What is 2+2?" })
});
const data1 = await res1.json();
// Response: { chatId: "abc-123", parentId: "xyz-789", ... }

// Continue conversation
const res2 = await fetch('/api/chat', {
  method: 'POST',
  body: JSON.stringify({ 
    message: "Now add 3?",
    chatId: data1.chatId,
    parentId: data1.parentId
  })
});

⁠🤖 Available Models

⁠Qwen 3.5

• qwen3.5-plus, qwen3.5-flash, qwen3.5-397b-a17b, qwen3.5-122b-a10b, qwen3.5-27b

⁠Qwen 3

• qwen3-max, qwen3-plus, qwen3-omni-flash, qwen3-coder-plus

⁠Vision Models

• qwen-vl-max, qwen-vl-plus, qwen2.5-vl-32b-instruct, qwen2.5-vl-7b-instruct

⁠Standard Models

• qwen-max, qwen-plus, qwen-turbo, qwen2.5-72b-instruct, qwq-32b

⁠🔑 API Key Authorization

By default, API is open. To enable authorization, add keys to src/Authorization.txt:

your-api-key-here

Then include header: Authorization: Bearer your-api-key-here

⁠🌐 Telegram Bot Integration

Enhanced fork includes Telegram bot for:

• Session upload (.zip/.7z archives)
• LLM chat assistant
• Real-time health monitoring
• Automatic session backups
• Graceful service restarts

Configure via environment variables:

• TELEGRAM_BOT_TOKEN - Bot token from @BotFather
• TELEGRAM_ALLOWED_USERS - Comma-separated user IDs (optional whitelist)

⁠📊 Environment Variables

⁠🗂️ Docker Volumes

⁠🔧 Account Management

⁠Interactive Menu

Run npm start without SKIP_ACCOUNT_MENU:

1. Add Account - Opens browser for Qwen login
2. Re-login - Refresh expired token
3. Start Proxy - Launch server (default)
4. Delete Account - Remove saved account

⁠Account Statuses

• OK - Active and valid
• WAIT - Rate limited (auto-resets after 24h)
• INVALID - Token expired, needs re-login

⁠Auto-Rotation

With multiple accounts:

• Round-robin selection
• Auto-switch on HTTP 429 (rate limit)
• Auto-switch on HTTP 401 (unauthorized)

⁠📝 API Reference

⁠POST /api/chat/completions (OpenAI Compatible)

{
  "model": "qwen-max-latest",
  "messages": [
    {"role": "system", "content": "You are a helpful assistant"},
    {"role": "user", "content": "Hello!"}
  ],
  "stream": false
}

Response:

{
  "id": "chatcmpl-123",
  "choices": [{"message": {"role": "assistant", "content": "Hi!"}}],
  "chatId": "abc-123",
  "parentId": "def-456"
}

⁠POST /api/chat (Native)

{
  "message": "Hello!",
  "model": "qwen-max-latest",
  "chatId": "abc-123",
  "parentId": "def-456"
}

⁠POST /api/files/upload

curl -X POST http://localhost:3264/api/files/upload -F "[email protected]"

Response:

{
  "success": true,
  "file": {
    "url": "https://cdn.qwenlm.ai/uploaded-image.jpg"
  }
}

⁠POST /api/images/generations

{
  "prompt": "Space ship on nebula background",
  "model": "qwen-image-plus",
  "n": 1,
  "size": "1024x1024"
}

⁠🔌 OpenWebUI Integration

1. Base URL: http://localhost:3264/api
2. API Key: Any (or leave empty if Authorization.txt is empty)
3. Streaming: Fully supported
4. Chat Isolation: Without conversation_id, proxy doesn't restore global context

⁠📚 Documentation

Full documentation available in repository:

• docs/ENV_SETUP.md - Environment configuration
• docs/TELEGRAM_BOT_FLOW.md - Telegram bot workflows
• docs/IMAGE_GENERATION.md - Image generation guide
• docs/OPENWEBUI_SETUP.md - OpenWebUI integration
• docs/TOKEN_EXPIRY_CHECKING.md - Token expiry detection

⁠🛡️ Security

• Credentials never logged
• Proxy support for Telegram API (HTTP/HTTPS/SOCKS)
• User whitelist for Telegram bot
• Secure file permission management

⁠📦 Project Structure

├── index.js              # Entry point: Express server
├── main.py               # Python FastAPI alternative
├── src/
│   ├── api/              # API routes and handlers
│   ├── browser/          # Puppeteer browser automation
│   ├── logger/           # Winston logging
│   └── utils/            # Utilities (account setup, Telegram bot)
├── session/              # Tokens and accounts
├── logs/                 # Log files
└── examples/             # Usage examples (JS/Python)

⁠🤝 Contributing

This is an enhanced fork of https://github.com/y13sint/FreeQwenApi⁠

Key improvements:

• Telegram bot integration
• Session management via Telegram
• LLM chat assistant
• Enhanced logging (Winston)
• Archive auto-extraction
• Health monitoring
• Fault-tolerant operations

⁠📄 License

MIT License - see LICENSE file for details