ai-medialisted

# Media ## Purpose Generate images, videos, and audio using fal.ai models via MCP. Progressive quality pattern: iterate cheap, finalize expensive. Covers text-to-image, text/image-to-video, text-to-speech, and video-to-audio. ## When to Use - `image`: generating images from text prompts (thumbnails, hero images, insert shots) - `video`: creating videos from text or images (demos, b-roll, social clips) - `audio`: generating speech, music, or sound effects (voiceover, background music, SFX) ## Process ### Step 1 -- Gate Check (MCP Required) Verify the fal.ai MCP server is configured. If not available, inform the user and provide setup instructions: ```json "fal-ai": { "command": "npx", "args": ["-y", "fal-ai-mcp-server"], "env": { "FAL_KEY": "YOUR_FAL_KEY_HERE" } } ``` Get an API key at [fal.ai](https://fal.ai). ### Step 2 -- Estimate Cost Before generating, always check estimated cost: ``` estimate_cost(model_name: "fal-ai/...", input: {...}) ``` Inform the user of the estimate before proceeding with expensive generations (video especially). ### Step 3 -- ElevenLabs Gate Check Before using ElevenLabs, verify `ELEVENLABS_API_KEY` is set. If not, fall back to csm-1b or inform the user. ### Step 4 -- Generate with Progressive Quality Start with cheaper models for prompt iteration, then switch to production models for finals. ### Step 5 -- Deliver Provide the generated media with: - file path or URL - model used and parameters - cost incurred - suggestion