Fal.ai MCP Server

Add to your client

Copy the config for your MCP client and paste it into its config file.

{
  "mcpServers": {
    "fal-ai-mcp-server": {
      "command": "uvx",
      "args": [
        "--from",
        "fal-mcp-server",
        "fal-mcp"
      ],
      "env": {
        "FAL_KEY": "<your-fal-api-key>"
      }
    }
  }
}

Before you start

• Python 3.10 or higher (or Docker, or the `uv`/`uvx` package manager for zero-install)
• A fal.ai API key set as the FAL_KEY environment variable (get it from your fal.ai dashboard; a free tier is available)
• An MCP client such as Claude Desktop; HTTP/SSE mode additionally needs `mcp-remote` or a compatible web client

About Fal.ai MCP Server

Tools & capabilities (18)

When to use it

• Use it when you want one MCP connection that spans image, video, and music generation instead of separate servers per modality.
• Use it when you need image editing primitives (background removal, upscaling, inpainting, resizing, compositing) inside an agent.
• Use it when you want text-to-video or image-to-video from inside Claude or another MCP client.
• Use it when you need to discover or compare fal.ai models and their pricing before committing to a job.
• Use it when you want HTTP/SSE access so a web-based client can reach the server.
• Use it when you want zero-install via uvx or a containerized deployment via Docker.

Quick setup

1. 1Get a fal.ai API key from your fal.ai dashboard.
2. 2Install via `uvx --from fal-mcp-server fal-mcp`, `pip install fal-mcp-server`, or the Docker image.
3. 3Add the server to your MCP client config with `fal-mcp` as the command (or `mcp-remote` for the HTTP/Docker variant).
4. 4Set FAL_KEY in the env block; optionally set FAL_MCP_TRANSPORT for stdio/http/dual mode.
5. 5Restart the MCP client so it picks up the server.
6. 6Call list_models or generate_image to verify the connection.

Fal.ai MCP Server FAQ