A 12-second sting scored from a product still: image_url on Sume music
Sume music takes an optional image_url and no duration field. Put "a 12-second track" in the prompt; the job is a flat $0.125 whatever the image or the length.

To score a 12-second sting from a product still on Sume, send prompt and a public HTTPS image_url to the music route, and write the length into the prompt. Sume rejects duration and duration_seconds on music requests, so "a 12-second track" in the text is the control. The price is a fixed $0.125 per generation, regardless of the image or the length you ask for.
This is from the Sume Music 1.0 and Music Router docs and the API catalog (read 2026-10-09). The router is the recommended route for new integrations; Music 1.0 requests already resolve through it.
What the request looks like
The prompt is 1 to 5,000 characters and should say what the brief needs: emotion, genre, tempo as a number, key, two to four instruments, an arc with one named moment, and the era. For a sting, say "12-second track" and name the moment, for example "hit at 0:09". End with "Instrumental, no vocals."
Exclusions go in the positive prompt: a non-empty negative_prompt returns HTTP 400 with negative_prompt_unsupported.
| Field | Rule |
|---|---|
| prompt | 1 to 5,000 characters, required |
| image_url | Optional public HTTPS image, or null to clear |
| duration, duration_seconds | Rejected; ask in the prompt |
| negative_prompt | Omit or send an empty string |
| Price | $0.125 flat per generation |
What the image does and does not do
The docs list image_url as visual conditioning for the music brief. They do not document it as a timing control, so do not expect it to follow cuts in your video. If the sting must land on a frame, check the generated audio and use a trim or a timeline offset.
Four candidates from four stills cost 4 x $0.125 = $0.50. Judge them by listening, since the length in the prompt is a request and not a guaranteed value.
After you pick one
The audio artifact is a media.sume.com URL that you can use as the soundtrack of a Timeline 1.0 render, with gain_db, loop and duck_db if there is a voice track. For a clean 12-second cut from a longer result, trim the audio in the timeline instead of asking for another generation.
Writing the brief
A workable 12-second brief reads: "Tight, confident product sting, 124 BPM, F major. Plucked synth and a dry kick, a bright bell hit at 0:09, clean tail. A 12-second track. Instrumental, no vocals." It names the tempo, key, two instruments, one moment and the length.
Because Sume cannot guarantee the length, plan to trim. A 14-second result is easy to cut to 12 seconds in a timeline, while a 9-second result would need another generation.
Music 1.0 routes keep working and keep job.model as sume/music-1.0, while the work runs through the router. job.request.routed_model names the engine that ran, such as lyria-3.5, so you can see what produced each candidate.
Sources
Related posts
More in Media tools
- One ad for X and Snapchat: no shared aspect ratio, render two masters
X lists 16:9 or 1:1 for video ads and Snapchat lists 9:16, so they share no ratio. A 140-second length fits both. Two Sume renders cost $0.60.
- Pause-cut with Timeline alone: 14 slots from one source, no trims
One Timeline render can cut 14 pauses from a single clip with source_in slots and audio.parts: $0.11 total, versus $0.39 with 14 trim jobs first.
- Pause cuts from STT words: a 0.35 s gap rule and the 0.2 s trim floor
Turn STT word times into keep-ranges: cut gaps over 0.35 s, drop slivers under the 0.2 s trim minimum, and price 8 trim jobs at $0.02 each.
- HEVC phone clip for LinkedIn: exact trim re-encodes, keyframe copies
LinkedIn lists H.264 or VP8. Sume's exact trim re-encodes to libx264, while keyframe trim copies the stream, so a phone HEVC source stays HEVC with keyframe.
Written by Sume