Wan2GP

You can run this MCP server to give your AI assistants access to Wan2GP video generation capabilities. It bridges your MCP client with a Wan2GP Gradio server, enabling text-to-video, image-to-video, model discovery, and health checks through a stable, scriptable interface.

How to use

You will integrate the Wan2GP MCP Server with your MCP client to submit video generation tasks from conversations. Start the local MCP server, then configure your MCP client (such as Claude Desktop) to communicate with it. From there you can request videos generated from text prompts, animate static images, view the current generation queue, list available models and LoRA adapters, and perform health checks to ensure Wan2GP is reachable.

How to install

Prerequisites you need before installation.

1. Wan2GP Gradio server must be running on your host.

2. Python 3.10 or newer.

3. Claude Desktop is optional for MCP integration.

Step by step installation and setup

# 1. Clone the MCP server repository
cd /path/to/your/projects
git clone <this-repo> wan2gp-mcp-server
cd wan2gp-mcp-server

# 2. Install dependencies
pip install -r requirements.txt

# 3. Configure the MCP server URL (example shown via config.json)

{
  "wan2gp_url": "http://localhost:7860",
  "timeout": 300,
  "max_concurrent_tasks": 3,
  "default_model": "wan",
  "default_resolution": "1280x720",
  "output_directory": "./output"
}

# 4. Run the MCP server directly for standalone testing
python wan2gp_mcp_server.py

"],

## Configuration and runtime parameters

The MCP server is configured to point to your Wan2GP Gradio instance and to control runtime behavior. Use the following settings to customize how tasks are submitted and processed.

{ "wan2gp_url": "http://localhost:7860", "timeout": 300, "max_concurrent_tasks": 3, "default_model": "wan", "default_resolution": "1280x720", "output_directory": "./output" }

"],

Integration with Claude Desktop

To connect Claude Desktop to this MCP server, add a new MCP server entry pointing to the local Python process that runs the MCP server.

{
  "mcpServers": {
    "wan2gp": {
      "command": "python",
      "args": [
        "/absolute/path/to/wan2gp-mcp-server/wan2gp_mcp_server.py"
      ],
      "env": {
        "WAN2GP_URL": "http://localhost:7860",
        "WAN2GP_TIMEOUT": "300",
        "LOG_LEVEL": "INFO"
      }
    }
  }
}

"],

## Usage examples

After integration, you can request video generation from natural language prompts or image prompts. For example, you can ask to generate a short video of a specific scene or to animate a given image, then pick a model and resolution to suit your needs.

## MCP Tools

The server provides the following tools for creating and managing videos.

## Troubleshooting

If you cannot connect to Wan2GP, ensure the Gradio server is running and the URL matches your configuration. Verify the port (default 7860). If you see a Flash-Attention version conflict, resolve Wan2GP dependencies in its environment. For MCP tool visibility in Claude Desktop, check logs and ensure absolute paths are used in the config.

## Notes

The MCP server enables practical workflows by bridging conversational prompts with Wan2GP’s video generation capabilities, while providing queue management and status checks to monitor progress.

## Available tools

### generate\_text\_to\_video

Create a video from a descriptive text prompt with configurable resolution, length, and quality settings.

### generate\_image\_to\_video

Animate a static image based on a text prompt, with controllable motion and duration.

### health\_check

Check if the Wan2GP server is running and reachable, returning status and version.

### list\_models

Return the list of available video generation models supported by Wan2GP.

### list\_loras

Return the list of available LoRA adapters for Wan2GP models.

### get\_queue

Retrieve the current queue of ongoing and scheduled video generation tasks.

### cancel\_task

Cancel a specific queued generation task by its identifier.