Vidu

You run a dedicated MCP server that uses Vidu’s AI models to generate videos from images. This server exposes configurable features, tracks generation progress, and supports asynchronous notifications via a callback URL so you can integrate video creation into your workflows smoothly.

How to use

You interact with the Vidu MCP Server through an MCP client. You can submit requests to convert a single image into a short video, check the status of ongoing tasks, and upload images for use with the Vidu API. The server supports multiple models and time/resolution constraints, and can attach background music for 4-second videos. You can also provide a callback URL to receive updates when the generation status changes.

How to install

Prerequisites you need before installing: Node.js version 14 or newer, a Vidu API key, and TypeScript for development.

Option A: Smithery installation (auto-install for Claude Desktop)

npx -y @smithery/cli install @el-el-san/vidu-mcp-server --client claude

Option B: Gemini CLI configuration requires a server entry in your Gemini settings with the runtime command and environment variable for the API key.

{
  "mcpServers": {
    "vidu": {
      "command": "node",
      "args": [
        "your_path/vidu-mcp-server/build/index.js"
      ],
      "env": {
        "VIDU_API_KEY": "your_api_key_here"
      }
    }
  }
}

Option C: Manual installation

1. Clone the repository and enter the directory

git clone https://github.com/el-el-san/vidu-mcp-server.git
cd vidu-mcp-server

2. Install dependencies

npm install

3. Create the environment file with your API key

VIDU_API_KEY=your_api_key_here

Usage notes for Gemini CLI and local development

To build TypeScript code for Gemini CLI usage, run the build step and then configure Gemini to load the MCP server. Restart the Gemini CLI after changes to load the MCP.

Configuration and end-to-end workflow

The server exposes a stdio-based runtime command that you should run to start the MCP server locally. The command uses Node to execute the built MCP index, and you pass your Vidu API key via an environment variable.

Available tools

image_to_video

Converts a still image to a video using the selected model and parameters such as duration, resolution, movement amplitude, and optional prompt and BGM.

check_status

Checks the progress of a video generation task and retrieves credit usage information for the ongoing job.

upload_image

Uploads an image to the Vidu API for use in video generation, supporting image types and size limits.