Qwen API

OpenAI-compatible API endpoints for Qwen AI

Features • Quick Start • Supported Endpoints • OpenAPI Docs • Usage Examples • License

🌟 Overview

Qwen API Proxy acts as a bridge between Qwen AI's proprietary API and the widely-adopted OpenAI API format. This allows developers to seamlessly integrate Qwen's advanced AI capabilities into their applications using familiar OpenAI-compatible endpoints.

Note: This is an unofficial proxy and not affiliated with Alibaba Cloud or Qwen AI.

📘 OpenAPI Docs

API documentation : https://qwen-api.readme.io/

Spec file: qwen.json (OpenAPI 3.1.0)
What it is: OpenAPI-ready API documentation covering all endpoints, OpenAI-compatible request/response shapes, security, and examples.
How to use:
- Import qwen.json into Swagger UI, Redocly, Postman, Bruno, or Insomnia.
- Generate typed clients with your preferred tool (e.g., openapi-generator, orval).
Servers: Defaults to https://qwen.aikit.club; you can change the host variable or edit the server URL after import.

🚀 Key Features

Feature	Description
🔁 OpenAI Compatibility	Drop-in replacement for OpenAI API calls
💬 Chat Completions	Text-based conversations with all Qwen models
🎨 Image Generation	Create stunning images from text prompts
✏️ Image Editing	Modify existing images with text instructions
🎬 Video Generation	Transform text into video content
🔬 Deep Research	Comprehensive research with web search and citations
👨🏻‍💻 Web Development	Generate interactive web components and UI elements
🏗️ Full-Stack Apps	Complete application development from frontend to backend
📊 Slides Generation	Create professional presentations with AI-generated slides
🔍 Web Search	Enable web search capabilities in conversations
🧠 Thinking Mode	Activate reasoning mode for complex problem solving
👁️ Vision Support	Analyze images, PDFs, and visual content
📁 Multimodal Files	Support for image, audio, video, and document uploads
🌍 CORS Support	Full cross-origin resource sharing support
🔧 Custom Tool Calling	OpenAI-compatible function/tool calling for select models
⚡ Edge Performance	Lightning-fast global deployment via Cloudflare Workers

🛠️ Supported Endpoints

Endpoint	Method	Description
`/v1/validate`	GET/POST	Validate token
`/v1/refresh`	GET/POST	Refresh token
`/v1/models`	GET	List available models
`/v1/chat/completions`	POST	Chat completions
`/v1/images/generations`	POST	Generate images
`/v1/images/edits`	POST	Edit existing images
`/v1/videos/generations`	POST	Generate videos
`/v1/chats/delete`	DELETE/POST	Delete all chats

🧠 Model Capabilities

Model Name	👁️ Vision	💡 Reasoning	🌐 Web Search	🔧 Tool Calling
Qwen3.7-Max	✅	✅	✅	✅
Qwen3.7-Plus	✅	✅	✅	❌
Qwen3.6-Plus	✅	✅	✅	✅
Qwen3.6-Max-Preview	❌	✅	❌	✅
Qwen3.6-35B-A3B	✅	✅	✅	❌
Qwen3.6-27B	✅	✅	❌	❌
Qwen3.5-Plus	✅	✅	✅	✅
Qwen3.5-Max-Preview	✅	✅	✅	✅
Qwen3.5-Flash	✅	✅	✅	✅
Qwen3.5-397B-A17B	✅	✅	✅	✅
Qwen3.5-122B-A10B	✅	✅	✅	❌
Qwen3.5-35B-A3B	✅	✅	✅	❌
Qwen3.5-27B	✅	✅	✅	❌
Qwen3.5-Omni-Plus	✅	❌	✅	❌
Qwen3.5-Omni-Flash	✅	❌	✅	❌
Qwen3-Max	❌	✅	✅	✅
Qwen3-Coder	✅	❌	✅	✅
Qwen3-Coder-Flash	✅	❌	✅	❌
Qwen3-235B-A22B-2507	✅	✅	✅	✅
Qwen3-30B-A3B-2507	✅	✅	✅	❌
Qwen3-Omni-Flash	✅	✅	✅	❌
Qwen3-VL-235B-A22B	✅	✅	❌	❌
Qwen3-VL-32B	✅	✅	✅	❌
Qwen3-VL-30B-A3B	✅	✅	❌	❌
Qwen3-Next-80B-A3B	✅	✅	✅	❌
Qwen2.5-Max	✅	✅	✅	❌
Qwen2.5-Plus	✅	✅	✅	❌
Qwen2.5-Turbo	✅	✅	✅	❌
Qwen2.5-Coder-32B-Instruct	✅	✅	✅	❌
Qwen2.5-VL-32B-Instruct	✅	✅	✅	❌
Qwen2.5-Omni-7B	✅	❌	✅	❌
Qwen-Image	✅	❌	❌	❌
Qwen-Video	✅	❌	❌	❌
Qwen-Deep-Research	❌	✅	❌	❌
Qwen-Web-Dev	✅	✅	❌	❌
Qwen-Full-Stack	✅	✅	❌	❌
Qwen-Slides	✅	✅	❌	❌

🚀 Quick Start

Use the Public Instance

The public instance is available at: https://qwen.aikit.club

💡 Usage Examples

Authentication

The proxy requires a Bearer token containing Qwen credentials:

const headers = {
  Authorization: "Bearer YOUR_QWEN_ACCESS_TOKEN",
  "Content-Type": "application/json",
};

Temporary Free Token

For quick testing, you can use this temporary token until it expires.

Token:

eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJpZCI6ImFkMjAzOGFmLWIxYWYtNDY2ZC05ZmY4LTgwY2Q3N2Q4ZGM2MCIsImxhc3RfcGFzc3dvcmRfY2hhbmdlIjoxNzc3MDQ4NTk0LCJleHAiOjE3ODE5NTY2Mzl9.ENADm_6vYyqQtG4VAIMaA0QvODIC_savi5n4SmdG-uo

Expires: 2026-06-20 11:57:19 UTC (2026-06-20 17:27:19 IST, UTC+05:30)
Note: This token is for evaluation only and will stop working after the expiration time.

How to Get Your Token

To obtain your Qwen API token, follow these steps:

Visit Qwen Chat: Go to chat.qwen.ai and log in to your account
Run the Token Extractor: Copy and paste the following JavaScript code into your browser's developer console (press F12 → Console tab):

javascript: (function () {
  if (window.location.hostname !== "chat.qwen.ai") {
    alert("🚀 This code is for chat.qwen.ai");
    window.open("https://chat.qwen.ai", "_blank");
    return;
  }
  function getApiKeyData() {
    const token = localStorage.getItem("token");
    if (!token) {
      alert("❌ qwen access_token not found !!!");
      return null;
    }
    return token;
  }
  async function copyToClipboard(text) {
    try {
      await navigator.clipboard.writeText(text);
      return true;
    } catch (err) {
      console.error("❌ Failed to copy to clipboard:", err);
      const textarea = document.createElement("textarea");
      textarea.value = text;
      textarea.style.position = "fixed";
      textarea.style.opacity = "0";
      document.body.appendChild(textarea);
      textarea.focus();
      textarea.select();
      const success = document.execCommand("copy");
      document.body.removeChild(textarea);
      return success;
    }
  }
  const apiKeyData = getApiKeyData();
  if (!apiKeyData) return;
  copyToClipboard(apiKeyData).then((success) => {
    if (success) {
      alert("🔑 Qwen access_token copied to clipboard !!! 🎉");
    } else {
      prompt("🔰 Qwen access_token:", apiKeyData);
    }
  });
})();

Get Your Token: The script will automatically:
- Extract your access_token from localStorage
- Copy the access_token to your clipboard
Use the Token: The copied token is now ready to use as your Bearer token in API requests

Important Notes:

⚠️ This script only works on chat.qwen.ai - make sure you're logged in
🔒 Keep your token secure - it provides access to your Qwen account
🔄 You may need to regenerate the token periodically if it expires

Validate Token (from JS snippet)

Validate the access token produced by the browser JS snippet above.

curl -X POST https://qwen.aikit.club/validate \
  -H "Content-Type: application/json" \
  -d '{"token": "YOUR_QWEN_ACCESS_TOKEN"}'

Or via GET:

curl "https://qwen.aikit.club/validate?token=YOUR_QWEN_ACCESS_TOKEN"

Chat Completions

const response = await fetch("https://qwen.aikit.club/v1/chat/completions", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    model: "qwen-max-latest",
    messages: [{ role: "user", content: "Hello, how are you?" }],
    stream: false,
  }),
});

Image Generation

const response = await fetch("https://qwen.aikit.club/v1/images/generations", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    prompt: "A beautiful sunset over mountains",
    size: "1024x1024",
  }),
});

Image Editing

// Using FormData for file upload
const formData = new FormData();
formData.append("image", imageFile); // File object
formData.append("prompt", "Change the sky to a starry night");

const response = await fetch("https://qwen.aikit.club/v1/images/edits", {
  method: "POST",
  headers: {
    Authorization: "Bearer YOUR_QWEN_ACCESS_TOKEN",
  },
  body: formData,
});

// Or using JSON with image URL/base64
const response = await fetch("https://qwen.aikit.club/v1/images/edits", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    image: "https://example.com/image.jpg", // or base64 data URL
    prompt: "Add a rainbow in the background",
  }),
});

Web Search Mode

const response = await fetch("https://qwen.aikit.club/v1/chat/completions", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    model: "qwen-max-latest",
    messages: [{ role: "user", content: "What are the latest AI developments?" }],
    tools: [{ type: "web_search" }],
  }),
});

Thinking Mode

const response = await fetch("https://qwen.aikit.club/v1/chat/completions", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    model: "qwen-max-latest",
    messages: [{ role: "user", content: "Solve this complex math problem: ..." }],
    enable_thinking: true,
    thinking_budget: 30000,
  }),
});

Custom Tool/Function Calling

Qwen's web API (chat.qwen.ai) does not natively support tool/function calling — unlike the official DashScope API, the web reverse API has no built-in tool calling capability. This proxy bridges that gap by implementing OpenAI-compatible tool calling via prompt engineering with XML-based parsing, enabling standard OpenAI tools and tool_choice parameters on the web API.

Supported Models: Qwen3.6-Plus, Qwen3.5-Plus, Qwen3.5-Flash, Qwen3.5-397B-A17B, Qwen3-Max, Qwen3-235B-A22B-2507

const response = await fetch("https://qwen.aikit.club/v1/chat/completions", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    model: "qwen3.6-plus",
    messages: [{ role: "user", content: "What's the weather in Tokyo?" }],
    tools: [
      {
        type: "function",
        function: {
          name: "get_weather",
          description: "Get the current weather for a location",
          parameters: {
            type: "object",
            properties: {
              location: { type: "string", description: "City name" },
            },
            required: ["location"],
          },
        },
      },
    ],
    tool_choice: "auto",
    stream: true,
  }),
});

Supported tool_choice values:

"auto" / "required" — Model must use at least one tool (first turn), decides freely on subsequent turns
"none" — No tools, plain text response only
{ "type": "function", "function": { "name": "xyz" } } — Force a specific tool

Tool calling still instructs Qwen to emit XML as the primary format: <tool_calls>[{"name":"tool_name","arguments":{}}]</tool_calls>. Internally, the proxy also accepts a few fallback formats from Qwen output and maps qwen-safe tool aliases back to the original client tool names. External clients should continue sending normal OpenAI tools, tool_choice, and messages; response tool_calls use the original names.

Code Generation (qwen3-coder-plus)

Note: qwen3-coder-plus supports Qwen Code — a coding agent that operates in digital environments and can issue function/tool calls. This API supports handling the function calls produced by the agent.

const response = await fetch("https://qwen.aikit.club/v1/chat/completions", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    model: "qwen3-coder-plus",
    tools: [{ type: "code" }],
    messages: [{ role: "user", content: "Write a JavaScript function to add two numbers" }],
    stream: true,
  }),
});

Video Generation

const response = await fetch("https://qwen.aikit.club/v1/videos/generations", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    prompt: "A cat playing with a ball of yarn in slow motion",
    size: "1280x720",
  }),
});

Deep Research

const response = await fetch("https://qwen.aikit.club/v1/chat/completions", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    model: "qwen-deep-research",
    messages: [
      {
        role: "user",
        content: "Research the latest developments in quantum computing",
      },
    ],
    stream: false,
  }),
});

Web Development (qwen-web-dev)

The qwen-web-dev model is specialized for frontend web development, creating interactive web components, HTML/CSS/JavaScript code, and providing live preview capabilities.

Features:

HTML/CSS/JavaScript code generation
Interactive UI components
Responsive design support
Real-time preview generation
Framework support: React, Vue, Vanilla JS, HTML5
Styling: Tailwind CSS, Bootstrap

const response = await fetch("https://qwen.aikit.club/v1/chat/completions", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    model: "qwen-web-dev",
    messages: [
      {
        role: "user",
        content:
          "Create a responsive navigation bar with a logo, menu items, and a mobile hamburger menu using HTML, CSS, and vanilla JavaScript",
      },
    ],
    stream: false,
  }),
});

Example Output: The model will generate complete, production-ready web components with:

Clean, semantic HTML structure
Modern CSS with responsive breakpoints
Vanilla JavaScript for interactivity
Mobile-first design approach
Accessibility considerations

Full-Stack Development (qwen-full-stack)

The qwen-full-stack model handles complete application development, from frontend to backend, database design, API development, and system architecture.

Features:

Frontend and backend code generation
Database schema design
RESTful and GraphQL API development
Authentication and authorization
Microservices architecture
Deployment-ready code
Multi-language support: JavaScript, TypeScript, Python, Java, Go, PHP
Frameworks: React, Vue, Angular, Node.js, Express, Django, Flask, Spring Boot

const response = await fetch("https://qwen.aikit.club/v1/chat/completions", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    model: "qwen-full-stack",
    messages: [
      {
        role: "user",
        content:
          "Create a complete REST API for a task management system with user authentication, CRUD operations for tasks, and a React frontend. Use Node.js/Express for the backend and MongoDB for the database.",
      },
    ],
    stream: false,
  }),
});

Example Full-Stack Application:

// Advanced example: Building a complete blog platform
const response = await fetch("https://qwen.aikit.club/v1/chat/completions", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    model: "qwen-full-stack",
    messages: [
      {
        role: "user",
        content: `Build a complete blog platform with the following requirements:

Backend (Node.js/Express):
- User authentication with JWT
- CRUD operations for blog posts
- Comment system
- Like/bookmark functionality
- Image upload support
- RESTful API endpoints

Frontend (React):
- Home page with post listings
- Post detail page with comments
- Create/Edit post interface
- User profile page
- Responsive design with Tailwind CSS

Database (MongoDB):
- User schema with authentication
- Post schema with relationships
- Comment schema
- Proper indexing for performance`,
      },
    ],
    stream: false,
  }),
});

Key Differences:

Feature	qwen-web-dev	qwen-full-stack
Focus	Frontend UI/UX	Complete application stack
Code Output	HTML, CSS, JavaScript	Frontend + Backend + Database
Use Cases	Web components, landing pages, UI elements	Complete apps, APIs, microservices
Complexity	Simple to moderate	Moderate to complex
Architecture	Client-side only	Full system architecture

Slides Generation (qwen-slides)

The qwen-slides model generates professional presentations with AI-generated slide images and a downloadable PDF. It researches the topic, creates content, and renders visual slides automatically.

Features:

AI-powered slide content generation
Web search for up-to-date information
Auto-generated slide images for each page
Downloadable PDF of the full presentation
Thinking summary with research insights

const response = await fetch("https://qwen.aikit.club/v1/chat/completions", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    model: "qwen-slides",
    messages: [
      {
        role: "user",
        content: "Create a presentation on the history and future of artificial intelligence",
      },
    ],
    stream: false,
  }),
});

Example Output: The response includes:

Reasoning content: Thinking summary with research insights
Text content: Written description of each slide
Slide images: Markdown image links for each generated slide
PDF link: A downloadable PDF of the complete presentation

// Streaming example with thinking enabled
const response = await fetch("https://qwen.aikit.club/v1/chat/completions", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    model: "qwen-slides",
    messages: [
      {
        role: "user",
        content:
          "Make a 10-slide pitch deck for a SaaS startup that provides AI-powered customer support",
      },
    ],
    stream: true,
    enable_thinking: true,
    thinking_budget: 30000,
  }),
});

Delete All Chats

// Using DELETE method
const response = await fetch("https://qwen.aikit.club/v1/chats/delete", {
  method: "DELETE", // GET and POST are also supported
  headers: headers,
});

📁 Multimodal File Support

The API supports various file formats for comprehensive multimodal interactions:

⚠️ Important Limitation: Multiple inputs of the same modality category are not supported. Image, Audio, and Video are considered the same category (media files), while Documents (PDF, TXT, etc.) are a separate category. You can combine different categories (e.g., image + PDF) but cannot combine files within the same category (e.g., image + video).

Supported File Types

Media Files (same category):
- Images: JPG, PNG, GIF, WebP (most common), BMP, TIFF, ICO, ICNS, JFIF, JP2
- Audio: MP3, WAV, M4A, AAC (most common), AMR
- Video: MP4, MOV, AVI, MKV (most common), WMV, FLV
Documents (separate category): PDF, TXT, MD (most common), DOC, DOCX, CSV, XLS, XLSX

💡 Tip: Bold formats are the most commonly used and recommended for best compatibility.

📏 File Limits

The following limits apply to multimodal file uploads:

File Type	Max Size (MB)	Max Count	Max Duration (seconds)
Images	10	5	-
Audio	100	1	180
Video	500	1	600
Documents	20	5	-
Default	20	-	-

📋 Summary: You can upload up to 5 images (10MB each), 1 audio file (100MB, 3 minutes), 1 video file (500MB, 10 minutes), or 5 documents (20MB each) per request.

✅ Valid Combinations

✅ Multiple images
✅ Image + PDF
✅ Audio + PDF
✅ Video + PDF
✅ Single image/audio/video only

❌ Invalid Combinations

❌ Image + Audio
❌ Image + Video
❌ Audio + Video
❌ Multiple videos
❌ Multiple audio files

Vision-Style Multimodal Chat

// Analyze any supported file type using standard chat completions
const response = await fetch("https://qwen.aikit.club/v1/chat/completions", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    model: "qwen-max-latest",
    messages: [
      {
        role: "user",
        content: [
          { type: "text", text: "What do you see in this image?" },
          {
            type: "image_url",
            image_url: {
              url: "https://download.samplelib.com/png/sample-hut-400x300.png",
              // or use base64: "data:image/jpeg;base64,..."
            },
          },
        ],
      },
    ],
  }),
});

Valid Multimodal Combination (Image + PDF)

// ✅ VALID: Combine different categories (Media + Document)
const response = await fetch("https://qwen.aikit.club/v1/chat/completions", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    model: "qwen-max-latest",
    messages: [
      {
        role: "user",
        content: [
          { type: "text", text: "Analyze this image and PDF document together" },
          {
            type: "image_url",
            image_url: { url: "https://download.samplelib.com/png/sample-hut-400x300.png" },
          },
          {
            type: "file_url",
            file_url: { url: "https://pdfobject.com/pdf/sample.pdf" },
          },
        ],
      },
    ],
  }),
});

❌ Invalid Combinations (Don't Do This)

// ❌ INVALID: Cannot combine image + video (same category)
const response = await fetch("https://qwen.aikit.club/v1/chat/completions", {
  method: "POST",
  headers: headers,
  body: JSON.stringify({
    model: "qwen-max-latest",
    messages: [
      {
        role: "user",
        content: [
          { type: "text", text: "This will not work properly" },
          {
            type: "image_url",
            image_url: { url: "https://download.samplelib.com/png/sample-hut-400x300.png" },
          },
          {
            type: "video_url",
            video_url: { url: "https://download.samplelib.com/mp4/sample-10s.mp4" },
          },
          // ❌ Cannot mix media files (image, audio, video)
        ],
      },
    ],
  }),
});

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.

Built with ❤️ by Tarun

_{If you find this project useful, please consider giving it a ⭐ on GitHub!}

Name		Name	Last commit message	Last commit date
Latest commit History 5 Commits
.github/workflows		.github/workflows
LICENSE		LICENSE
README.md		README.md
WORKFLOW.md		WORKFLOW.md
qwen.json		qwen.json

Folders and files

Latest commit

History

Repository files navigation

Qwen API

🌟 Overview

📘 OpenAPI Docs

API documentation : https://qwen-api.readme.io/

🚀 Key Features

🛠️ Supported Endpoints

🧠 Model Capabilities

🧠 Model Capabilities

🚀 Quick Start

Use the Public Instance

💡 Usage Examples

Authentication

Temporary Free Token

How to Get Your Token

Validate Token (from JS snippet)

Chat Completions

Image Generation

Image Editing

Web Search Mode

Thinking Mode

Custom Tool/Function Calling

Code Generation (qwen3-coder-plus)

Video Generation

Deep Research

Web Development (qwen-web-dev)

Full-Stack Development (qwen-full-stack)

Slides Generation (qwen-slides)

Delete All Chats

📁 Multimodal File Support

Supported File Types

📏 File Limits

✅ Valid Combinations

❌ Invalid Combinations

Vision-Style Multimodal Chat

Valid Multimodal Combination (Image + PDF)

❌ Invalid Combinations (Don't Do This)

📜 License

About

Topics

Resources

License

Uh oh!

Stars

Watchers

Forks

Contributors

Uh oh!