✨ Ask AI — suggest a design from a photo
POST/me/avatar/from-photo
Reads a photo with an on-infrastructure vision model
(@cf/meta/llama-3.2-11b-vision-instruct) and returns avataaars
options resembling it. The photo is transient — downscaled,
described, and discarded; never written to R2, never logged. Nothing is
saved; the console seeds the Design editor with the result. Rate-limited
hard (each call costs Workers AI neurons).
Request
Responses
- 200
- 415
- 422
- 429
- 502
- 503
Suggested avataaars options.
Unsupported type (GIF is not a portrait source).
No face detected.
Rate limited.
Generation failed.
AI not configured.