The talking-avatar endpoint produces a video of a still portrait speaking
either a generated script (TTS) or an existing audio clip. Optional
product placement fuses a product image into the same shot.
Endpoints
Modes
Create Parameters
Idempotency-Key is required on every POST /v1/talking-avatars.The 100-character minimum for script is enforced server-side — shorter
scripts are rejected with invalid_request.script and audio_url cannot both be supplied. If mode=script but you
only pass audio_url, the request is auto-promoted to mode=audio.
Pricing
- 3 credits per second, minimum 5 seconds (so 15 credits floor).
- + 5 credits when
product_placement=true.
Example: script mode
Example: audio mode
Example: with product placement
Completed Response
Possible Errors