Seedance 2.5 adds subtitles or music I did not ask for: fixes
Seed says Seedance 2.5 minimises stray subtitles and music. On Sume, generate_audio false drops the audio track; subtitles need prompt wording.

If a Seedance 2.5 clip has background music you did not want, send generate_audio: false to drop the generated audio track; if it has burned-in subtitles, Sume has no setting for that, so you have to steer it in the prompt. ByteDance says the model minimises "uncontrolled occurrences in subtitles and background music", which concedes that both can still appear.
The quote is from ByteDance Seed's Seedance 2.5 launch post (read 2026-10-02). The parameter is documented in Sume's Video generation: generate_audio is a boolean that "defaults to the model's audio capability", and seedance-2.5 reports audio support.
Which lever fixes which problem?
Audio is a separate track in the output; subtitles are drawn into the picture. Sume's request has one field for the first and none for the second.
| What you see | Lever on Sume | What the docs say |
|---|---|---|
| Background music or sound you do not want | generate_audio: false | Boolean; defaults to the model's audio capability |
| Burned-in subtitles | Prompt wording only | No subtitle parameter in the request list |
| Wanted music from a specific track | reference_audio_urls (Video Router) or an audio_url input reference (/v1/videos) | Audio references are honoured by Seedance 2.x |
| Wanted captions, controlled | Add them after generation | See the subtitles post below |
How do I turn the audio off?
Set generate_audio to false in the request. It is optional, so leaving it out gives the model's default. Whether it changes the price is a separate question, answered in Do reference images or audio change the price?.
curl -X POST https://api.sume.com/v1/videos \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: silent-clip-001" \
-d '{"model":"seedance-2.5","prompt":"A ceramic mug rotating on a plain grey table, no on-screen text","duration":6,"generate_audio":false}'Can the prompt stop subtitles?
It can help; nothing here guarantees it. A plain instruction such as "no on-screen text, no subtitles, no captions" costs nothing to add. Seed's wording means the model has been tuned to produce fewer stray subtitles, not that the behaviour is removed.
If a clip still has text, regenerate with a changed prompt and a new Idempotency-Key. Reusing the key with a changed body returns a conflict, and reusing it with the same body returns the original job rather than a new clip.
What if I want subtitles, on my terms?
Generate clean video, then add captions as a separate step so the text, font and timing are yours. How to add subtitles to a video from a transcript covers that with Sume's tools, and Add background music to a video covers the audio side.
Sources
Related posts
More in Models
- Seedance 2.5 timestamp-level editing: no timestamp field on Sume
ByteDance describes timestamp-level control in Seedance 2.5. Sume documents no timestamp parameter on /v1/videos. Here is what you can send instead.
- Sume Auto video has no 1:1 or 21:9: which model to pin instead
Auto in Sume's Videos panel offers 720p or 1080p, 16:9 or 9:16 and 3 to 10 seconds. For square, ultrawide or longer clips, pin Wan 3.0, H3 or Kling 3.0.
- Reference limits per Sume video model: images, clips, audio
How many reference images, videos and audio clips each Sume video model takes on /v1/videos: Seedance 12 total, Wan 10/5/5, H3 9/3/3, Omni 10 and 3.
- Swap the product in a UGC clip with Gemini Omni Flash 1.1 video edit
Keep a UGC clip that works and change only the product: Sume's video router edit mode for gemini-omni-flash-1.1 takes a video_url and a one-line prompt.
Written by Sume