fal-ai-medialisted

# fal.ai Media Generation ## MCP Setup Add to `~/.claude.json` (API key from fal.ai): ```json "fal-ai": { "command": "npx", "args": ["-y", "fal-ai-mcp-server"], "env": { "FAL_KEY": "YOUR_FAL_KEY_HERE" } } ``` MCP tools: `search` (find models by keyword), `find` (model details/parameters), `generate` (run a model), `result` / `status` (async job state), `cancel`, `estimate_cost`, `models` (list popular), `upload` (upload local files for use as inputs). ## Image Models | app_id | Use for | |--------|---------| | `fal-ai/nano-banana-2` | fast iteration, drafts, image editing | | `fal-ai/nano-banana-pro` | production quality, realism, typography | ``` generate( app_id: "fal-ai/nano-banana-2", input_data: { "prompt": "a futuristic cityscape at sunset, cyberpunk style", "image_size": "landscape_16_9", "num_images": 1, "seed": 42 } ) ``` Image parameters: | Param | Type | Values | Notes | |-------|------|--------|-------| | `prompt` | string | required | | | `image_size` | string | `square`, `portrait_4_3`, `landscape_16_9`, `portrait_16_9`, `landscape_4_3` | | | `num_images` | number | 1-4 | | | `seed` | number | any integer | reproducibility | | `guidance_scale` | number | 1-20 | higher = more literal prompt adherence | | `image_url` | string | URL | source image for editing/style transfer | Image editing (inpainting, outpainting, style transfer) with Nano Banana 2: `upload(file_path: "/path/to/image.png")` first, then pass the returned URL as