fal.ai MCP Server

An MCP (Model Context Protocol) server that gives Claude access to fal.ai generative AI models — images, videos, audio, 3D models, and more.

Built with FastMCP. Works with Claude Code and Claude Desktop.

Model Catalog

Image Generation (10 models)

Image Editing (4 models)

Video Generation (5 models)

Audio Generation (4 models)

3D Generation (2 models)

Utility (2 models)

Available Tools

Synchronous (block until result, best for fast models)

Asynchronous queue API (for long-running jobs — video, 3D, large batches)

The synchronous tools block while the model runs. For models that take longer than the MCP tool-call timeout (~60 s), use the queue API instead: submit returns immediately with a request_id, then poll and fetch separately.

Utilities

Job metadata (request_id → model_id, media_type, prefix, status, files) is persisted in jobs.json next to server.py, so poll_job and fetch_job only need the request_id.

Example flow:

submit_job(model="kling-v3-pro", arguments={"prompt":"a fox jumping","duration":"5"}, media_type="video", prefix="fox")
  → {"request_id": "019dd...", "status": "SUBMITTED"}
poll_job("019dd...")           → {"status": "IN_PROGRESS", ...}
poll_job("019dd...")           → {"status": "COMPLETED"}
fetch_job("019dd...")          → {"files": [{"local_path": "~/Downloads/fal-ai/videos/fox_..."}]}

Installation

Prerequisites

• Python 3.10+
• A fal.ai account with an API key

1. Get your fal.ai API key

Go to fal.ai Dashboard > Keys and create an API key.

2. Set up the server

git clone https://github.com/Szotasz/fal-ai-mcp-server.git
cd fal-ai-mcp-server
python -m venv .venv
source .venv/bin/activate   # On Windows: .venv\Scripts\activate
pip install -r requirements.txt

3. Set your API key

export FAL_KEY="your-api-key-here"

Usage with Claude Code

claude mcp add fal-ai \
  -e FAL_KEY=your-api-key-here \
  -- /path/to/fal-ai-mcp-server/.venv/bin/python /path/to/fal-ai-mcp-server/server.py

Usage with Claude Desktop

Add this to your claude_desktop_config.json:

{
  "mcpServers": {
    "fal-ai": {
      "command": "/path/to/fal-ai-mcp-server/.venv/bin/python",
      "args": ["/path/to/fal-ai-mcp-server/server.py"],
      "env": {
        "FAL_KEY": "your-api-key-here"
      }
    }
  }
}

Replace /path/to/fal-ai-mcp-server with the actual path where you cloned the repo.

Output

Generated files are automatically downloaded to ~/Downloads/fal-ai/ organized by type:

• images/ — generated and edited images
• videos/ — generated videos
• audio/ — speech, music, and sound effects
• 3d/ — 3D models
• other/ — everything else

License

MIT