Tongyi Wanx

🚀 Tongyi Wanxiang MCP Server

This is a TypeScript-based Model Context Protocol (MCP) server that specifically provides the text-to-image and text-to-video capabilities of Alibaba Cloud's Tongyi Wanxiang. Through the MCP protocol, this server allows large language models (LLMs) to directly call Tongyi Wanxiang's image and video generation APIs.

✨ Features

• Text-to-Image Integration: Connects to Alibaba Cloud's Tongyi Wanxiang text-to-image API, supporting high-quality AI image generation.
• Text-to-Video Integration: Connects to Alibaba Cloud's Tongyi Wanxiang text-to-video API, supporting high-quality AI video generation.
• Asynchronous Task Handling: Supports long-running image and video generation tasks, and retrieves the final results through asynchronous polling.
• MCP Protocol Support: Complies with the Model Context Protocol specification and can seamlessly collaborate with LLMs that support MCP.

📦 Installation

Environment Requirements

• Node.js >= 16.x
• npm >= 8.x or pnpm

How to Use

Take the Bailian platform as an example:

{
  "mcpServers": {
    "tongyi-wanxiang": {
      "command": "npx",
      "args": [
        "-y",
        "tongyi-wanx-mcp-server@latest"
      ],
      "env": {
        "DASHSCOPE_API_KEY": "<Your Tongyi Wanxiang API Key>"
      }
    }
  }
}

How to Develop

Install Dependencies

# Using npm
npm install

# Or using pnpm
pnpm install

Build and Run

# Build the project
npm run build
# Or
pnpm run build

# Run the server
npm start
# Or
pnpm start

# Run with debugging tools
npm run debug
# Or
pnpm run debug

💻 Usage Examples

Basic Usage

The server provides the following MCP tools:

1. Text-to-Image Generation (wanx-t2i-image-generation)

Starts an image generation task and returns the task ID. Parameters:

• prompt: The prompt for image generation.
• negative_prompt: Negative prompt (elements that you don't want to appear in the image).

Return:

• Task information containing the task_id.

2. Get Generation Result (wanx-t2i-image-generation-result)

Queries the image generation result through the task ID. Parameters:

• task_id: The unique identifier of the generation task.

Return:

• Image data or error information.

3. Text-to-Video Generation (wanx-t2v-video-generation)

Starts a video generation task and returns the task ID. Parameters:

• prompt: The prompt for video generation.
• negative_prompt: Negative prompt (elements that you don't want to appear in the video).
• duration: The duration of the video (optional, default is 10 seconds).

Return:

• Task information containing the task_id.

4. Get Generation Result (wanx-t2v-video-generation-result)

Queries the video generation result through the task ID. Parameters:

• task_id: The unique identifier of the generation task.

Return:

• Video data or error information.

🔧 Technical Details

Project Structure

your-project
├── src/
│   ├── mcp-server.ts
│   └── utils/
│       └── config.ts
└── package.json

Configuration Instructions

Environment Variables

• DASHSCOPE_API_KEY: The access key for Tongyi Wanxiang API.
• PORT: The port that the server listens on (default is 3000).

Example Configuration File (config.ts):

export const CONFIG = {
  api_key: process.env.DASHSCOPE_API_KEY,
  port: parseInt(process.env.PORT) || 3000,
}

📄 Notes

1. Please ensure that you have obtained valid access permissions and keys for the Tongyi Wanxiang API.
2. Image generation is an asynchronous process and may take several seconds to tens of seconds.
3. Video generation is a time-consuming process and may take several minutes to more than ten minutes.
4. Video generation status queries may fail multiple times. The system will automatically retry, so please be patient.
5. Please reasonably set the polling interval and the maximum number of retries to suit your usage scenario.
6. For video generation tasks, it is recommended to increase the maximum number of retries and the polling interval.

📚 Documentation

• Tongyi Wanxiang API Documentation
• Model Context Protocol (MCP) Specification