MiniMax H3 Talk

Drive mouth timing and delivery from an uploaded speech clip via H3 Ref2VA. Upload a face (first_frame) plus 2–15s of talking audio; duration snaps to the clip.

Deploy & run Source

RTX 5090 scale to zero WebM AV1 + stereo audioMiniMax H3audio-drivenRef2VAtalking

Canonical example

Input and output, side by side

Upload audio + portrait · Ref2VA mouth sync

01 · input

reference image

driving audio

prompt

Medium close-up of the person in the reference portrait speaking warmly to camera; natural head motion, soft studio key light, shallow depth of field, mouth and expression follow the uploaded speech exactly.

aspect ratio

9:16

size

balanced

duration

steps

seed

31008

02 · WebM AV1 + stereo audio

Run the same recipe everywhere

curl -sX POST https://app.nz/api/cogs/run \
  -H "Authorization: Bearer $APP_NZ_KEY" \
  -H "Content-Type: application/json" \
  -d '{"template":"minimax-h3","input":{"prompt":"Medium close-up of the person in the reference portrait speaking warmly to camera; natural head motion, soft studio key light, shallow depth of field, mouth and expression follow the uploaded speech exactly.","aspect_ratio":"9:16","size":"balanced","duration":5,"steps":20,"seed":31008,"structured_prompt":true,"include_audio":true,"output_codec":"webm-av1","first_frame":"/static/cogs/h3/talking-portrait.png","audio":"/static/cogs/h3/talking-hello.wav"}}'