@genvr/mcp-server

GenVR MCP Server

Model Context Protocol (MCP) server for GenVR APIs, enabling seamless integration with Cursor IDE and other MCP-compatible tools.

Installation

Option 1: Use directly with npx (Recommended)

No installation needed! Just configure in your .cursor/mcp.json:

{
  "mcpServers": {
    "GenVR": {
      "command": "npx",
      "args": [
        "-y",
        "@genvr/mcp-server"
      ],
      "env": {
        "GENVR_USER_ID": "your_user_id_here",
        "GENVR_ACCESS_TOKEN": "your_access_token_here"
      }
    }
  }
}

Option 2: Install globally

npm install -g @genvr/mcp-server

Then configure in .cursor/mcp.json:

{
  "mcpServers": {
    "GenVR": {
      "command": "genvr-mcp",
      "env": {
        "GENVR_USER_ID": "your_user_id_here",
        "GENVR_ACCESS_TOKEN": "your_access_token_here"
      }
    }
  }
}

Configuration

• Get your GenVR API credentials from https://app.genvrresearch.com/teams

• Create .cursor/mcp.json in your project root with the configuration above

• Enable MCP in Cursor:

  • Open Settings (Cmd/Ctrl + ,)
  • Search for "MCP"
  • Enable "Model Context Protocol"

• Restart Cursor

Available Tools

generate_image

Generate images using GenVR's image generation models.

Parameters:

• prompt (required): Text description of the image to generate
• model (optional): Model name (default: "flux_dev")
  • Options: flux_dev, flux_schnell, sdxl_lightning, etc.
• width (optional): Image width in pixels (default: 1024)
• height (optional): Image height in pixels (default: 1024)
• num_inference_steps (optional): Number of inference steps (default: 50)

Example:

Generate a photorealistic image of a sunset over mountains

generate_video

Generate videos using GenVR's video generation models.

Parameters:

• prompt (required): Text description of the video to generate
• model (optional): Model name (default: "kling2_5_t2v")
  • Options: kling2_5_t2v, kling_i2v, runway_gen3, etc.
• duration (optional): Video duration in seconds (default: 5)

Example:

Create a 10-second video of a cat playing with a ball

generate_audio

Generate audio/music using GenVR's audio generation models.

Parameters:

• prompt (required): Text description of the audio to generate
• model (optional): Model name (default: "dia")
  • Options: dia, sonauto_text2music, suno, etc.
• duration (optional): Audio duration in seconds (default: 10)

Example:

Generate upbeat electronic music for a video game

genvr_custom

Call any GenVR model with custom parameters for advanced use cases.

Parameters:

• category (required): Model category (imagegen, videogen, audiogen, 3dgen, etc.)
• subcategory (required): Specific model name
• parameters (required): Model-specific parameters as JSON object

Example:

{
  "category": "imagegen",
  "subcategory": "flux_dev",
  "parameters": {
    "prompt": "A beautiful landscape",
    "width": 1920,
    "height": 1080
  }
}

Usage in Cursor

Once configured, you can use natural language in Cursor's AI chat:

@GenVR Generate a photorealistic image of a futuristic city at night

@GenVR Create a 5-second video of ocean waves

@GenVR Generate background music for a meditation app

The MCP server will automatically:

• Call the appropriate GenVR API
• Poll for task completion
• Return the generated content URLs

Development

To build from source:

npm install
npm run build
npm start

Environment Variables

• GENVR_USER_ID: Your GenVR user ID (required)
• GENVR_ACCESS_TOKEN: Your GenVR access token (required)
• GENVR_API_BASE: API base URL (optional, defaults to https://api.genvrresearch.com/api/v1)

Troubleshooting

"No tools, prompts or resources" message

• Ensure .cursor/mcp.json is in the correct location
• Verify MCP is enabled in Cursor settings
• Check that credentials are correct (no extra spaces)
• Restart Cursor completely

MCP server not loading

• Ensure Node.js 18+ is installed
• Check JSON syntax in .cursor/mcp.json
• View Cursor's output panel for error messages

Support

• API Documentation: https://api.genvrresearch.com
• Get API Key: https://app.genvrresearch.com/teams
• Report Issues: [GitHub Issues]

License

MIT