Lip-sync a saved avatar with H3 Max: avatar_handle, not image_url

To lip-sync a saved avatar with MiniMax H3 Max, post to POST /v1/minimax/h3-max/lip-sync with avatar_handle instead of image_url, a Sume-hosted audio_url, and a duration_seconds between 5 and 14.8. A 10-second line at 768p reserves $1.00 on Sume. The body is the same as VEED Fabric 1.0: exactly one visual source, either image_url or an avatar id or handle, never both. Sume lists the route in its models docs.

Why use the handle

A handle means the same face in every clip without hosting a still. Use a handle whose avatar job has finished. The audio must live on Sume's media host and be at most 10 MB, so run speech through Sume TTS first or upload through Sume's media inputs.

• resolution is 480p, 768p (default) or 1080p; there is no 2K.
• speed_tier is accepted and ignored.
• The API rejects model and provider endpoint fields: the URL is the model.

Price at each resolution

Sume's rate is the fal list price times 1.25. fal's own page lists $0.05, $0.08 and $0.16 per second at 480p, 768p and 1080p; Sume's per-second rates are $0.0625, $0.10 and $0.20.

Request

Replace the audio URL with your own Sume-hosted file.

curl -X POST https://api.sume.com/v1/minimax/h3-max/lip-sync \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: h3-handle-001" \
  -d '{
    "avatar_handle": "support_host",
    "audio_url": "https://media.sume.com/audio/line.mp3",
    "duration_seconds": 10,
    "resolution": "768p"
  }'

curl https://api.sume.com/v1/jobs/JOB_ID/status -H "Authorization: Bearer $SUME_API_KEY"