What Your Agent Can Do

Model types

  • Image: "Make a 1:1 product photo of a white ceramic mug with Muapi." Includes editing, upscaling and background removal.
  • Video: text-to-video, image-to-video, video editing and lipsync.
  • Music and speech: "Generate a 30 second lo-fi background track." / "Turn README.md into a spoken intro."
  • 3D: text-to-3D and image-to-3D.
  • LLM text and data tools: language models, SEO and data endpoints, OCR, transcription and moderation.

What happens when you ask

  1. Searches the live model list for matches.
  2. Fetches the model's parameters, valid values, defaults and price.
  3. Estimates the cost before spending credits on anything non-trivial.
  4. Uploads input files from your machine and supplies hosted URLs.
  5. Submits the task and polls until it completes.
  6. Downloads the outputs to the path you specified.

Writing strong requests

Include:

  • Output destination, such as ./public/hero.png.
  • A specific model, if you have a preference. Otherwise the agent states which model it picked.
  • Key parameters such as aspect ratio, resolution or duration.
  • Integration asks, if you want code that calls the API rather than a one-off file.

Limits

  • Tasks use Muapi credits. Check usage in the dashboard.
  • Generated files are kept for 30 days. The agent downloads results immediately so nothing is lost.
  • Input uploads: images up to 10MB, videos up to 50MB.
  • Social publishing needs a one-time browser OAuth approval from you before an agent can post.