AI sound effects for video: what Sume lists and what it does not

Sume's docs do not currently list a dedicated AI sound-effects generator. Sound for a video comes three ways: a video model that generates audio with the clip, a music track from the Music Router, or audio you already have, mixed in Timeline 1.0. If you need a one-shot effect like a door slam as a standalone file, Sume does not list an endpoint for it.

From the Video generation, Music Router and Timeline 1.0 docs, read 2026-09-29.

Which Sume routes make sound?

How do I get sound with a generated clip?

Check each model's generate_audio flag in the catalog. The request field generate_audio defaults to the model's audio capability, so set it to true and describe the sounds in the prompt: the docs list seedance-2 with audio, and minimax-h3-max with native stereo audio.

curl -X POST https://api.sume.com/v1/videos \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: sfx-clip-001" \
  -d '{
    "model": "seedance-2",
    "prompt": "Rain on a tin roof, a kettle whistles in the kitchen, close camera.",
    "duration": 5,
    "resolution": "720p",
    "aspect_ratio": "16:9",
    "generate_audio": true
  }'

Can I mix my own effects?

Yes, with files you already have. Import them to media.sume.com, join or trim with Timeline audio, and set a bed as the render's soundtrack. That is mixing existing sound, not generating new effects.