Can a Sume avatar speak my own recording? Voice types explained

The Avatar 1.0 talking-video docs describe speech as written text: a top-level script, or ordered video_inputs where each scene's voice block is either type: "text" or type: "silence". They do not document an audio-file input for this route, so plan on text in and a spoken avatar out (Generate avatar video).

If your workflow depends on a specific recording, check the live OpenAPI before you build around it, and do not assume the route accepts one.

What each voice type means

A scene list with a pause

The pause is a real scene with a duration, so it counts toward the 4 to 60 second window.

{
  "avatar_handle": "product_host",
  "quality": "plus",
  "video_inputs": [
    { "id": "hook", "voice": { "type": "text", "script": "Wait, one selfie became a whole video?", "duration": 3 } },
    { "id": "beat", "voice": { "type": "silence", "duration": 2 } },
    { "id": "cta",  "voice": { "type": "text", "input_text": "Pick a template and drop in your photo.", "duration": 4 } }
  ]
}

Cost of that plan

Nine seconds in total, billed at the tier rate with no product image.

Other constraints

The current execution supports one resolved avatar for each final video, and the scene backgrounds are expected to resolve to one shared scene. Total planned duration must stay in the 4 to 60 second window.