Generate an image locally
Generate an original image from a visual brief with your chosen local image model, for an article hero, thumbnail or illustration
Runs MFLUX on your Apple Silicon Mac through its public, MIT-licensed MCP server. Uses your computer for inference. The selected model is explicit; no paid provider or Trillion inference fallback. First use downloads that model. Returns the completed PNG in your files.
What it asks for
- brief (Needed, words): Describe the picture or requested edit. Include the scene and visual style; never copy a page title or subtitle into the image.
- model (Needed, words): The local model name you chose, from the available local image models. Its weight license must permit this use.
- steps (Optional, number):
- seed (Optional, number):
- quantize (Optional, one of):
- width (Optional, number):
- height (Optional, number):
Use it from your AI
Connect the AI you already use to Trillion once. Then ask in your own words; your AI finds this tool by what it does and runs it.
Say to your AI: “Use Trillion to generate an original image from a visual brief with your chosen local image model, for an article hero, thumbnail or illustration”
What your AI calls: run_tool {"tool_id":"tool:generate-local-image"}