Video

MiniMax H3 prompts for shots and sound

Write visual action, speech, and ambient sound on a shared timeline.

Break the clip into time ranges

Give each range a clear camera framing and one main action. Include a final frame that can be checked against the brief.

Separate sound from picture

Write dialogue verbatim and label background sounds. Leave enough time for the line to be spoken.

Check supported references

Select the exact H3 variant and read its reference roles and parameters before requesting a quote.

Example request

curl "https://api.apimodel.store/v1/quotes" \
  -H "Authorization: Bearer $MODELPORT_API_KEY" \
  -H "Content-Type: application/json" \
  --data '{"model":"minimax-h3","prompt":"A clerk opens a wooden cafe window. The clerk says: The rain has stopped.","parameters":{"duration":8,"resolution":"768P","aspect_ratio":"16:9","generate_audio":true,"n":1}}'

Test prompt

0–4 seconds: fixed medium shot as a café worker opens a wooden window. 4–8 seconds: cut closer as they say, “The rain stopped.” Distant traffic and a soft wooden creak. No subtitles.

Models in this guide

Related guides and tools

Sources