| name | generate-asset |
| description | Generate a single media asset — image, video, audio, or 3D. Use when the user wants one output from a clear prompt. Calls muapi_select to pick the best model, then muapi_generate. |
Generate Asset
Use this skill for single-asset generation tasks.
Step 1 — Discover
Call muapi_select with the user's intent and kind:
muapi_select(intent="<user's prompt>", kind="<image|video|audio|3d>", tier="<best|balanced|fast>", limit=3)
Pick the top-ranked model from the result.
Step 2 — Generate
Call muapi_generate with the chosen model:
muapi_generate(
prompt="<refined prompt>",
kind="<kind>",
model="<name from muapi_select>",
tier="<best|balanced|fast>",
)
For edits, enhancements, image-to-video, or lipsync — pass input_asset_url too.
Step 3 — Return
Report the asset URL to the user. Include: model used, kind, and any relevant parameters.