mcp-fal-ai-image

The MCP fal.ai Image Server lets you generate images from text prompts directly from your IDE using fal.ai models. It wraps the fal.ai image generation capabilities in an MCP-compliant server so you can invoke a generate-image tool from your editor and receive image results locally. The server supports key image-generation parameters such as prompt, model ID (e.g., fal.ai models like fal-ai/kolors), number of images, image size, inference steps, guidance scale, and a safety checker toggle. Images are saved to a local directory and the tool response includes their file paths, making it easy to drop generated visuals into your project. You can configure your API key via FAL_KEY to authenticate with fal.ai when making requests. To use it in your IDE, select the generate-image tool, provide a prompt (and optional model and parameters), and inspect the returned file paths for the generated images.

Prerequisites: - Node.js 18+ with npm installed. - Access to fal.ai API key.

Installation steps:

1. Install Node.js 18+ if you don’t have it:

   • macOS (Homebrew): brew install node@18
   • Ubuntu/Debian: curl -fsSL https://deb.nodesource.com/setup_18.x | bash - apt-get install -y nodejs
   • Windows: download and install from https://nodejs.org/

2. Ensure npm is available:

   node -v npm -v

3. Run the MCP server via npx (as configured in the readme example):

   npx -y mcp-fal-ai-image

   This will start the MCP server using the provided configuration. Make sure to set your fal.ai API key in the environment:

   FAL_KEY=YOUR-FAL-AI-API-KEY

4. Verify the server is accessible by triggering a generate-image request from your IDE or via the MCP client tools.

5. Optional: Set up a global npm script or use your editor’s MCP integration to simplify running the server.

Tips and common issues:

• Ensure the FAL_KEY environment variable is set to your fal.ai API key before starting the server.
• If npx is not found, install Node.js 18+ and npm as described above.
• The images are saved to a local directory; check the MCP response for the file paths.
• If a model ID is not supported by fal.ai, ensure you’re using a valid fal.ai model from their catalog and double-check the model string in your prompt.
• You can override the default output directory via the FAL_IMAGES_OUTPUT_DIR environment variable to customize where images are saved.
• For troubleshooting, consult the MCP client logs in your IDE and verify network access to fal.ai endpoints.