Restyle your own footage with the Sume restyle Format
sume-restyle keeps the motion, timing and cuts of a clip you attach, rewrites its look chunk by chunk, and reassembles it with your original audio.
sume-restyle is the catalog Format for changing how footage you already shot looks without changing what happens in it. Its brief keeps the clip's motion, timing and cuts, rewrites the look chunk by chunk, and reassembles the result on the timeline with your original audio. Call it at the sume handle with the clip and a packshot or look reference attached.
What stays and what changes
The brief names Seedance 2.5 reference-to-video for the chunked rewrite, then the timeline for reassembly, so the Format is a recipe over tools Sume already lists. You do not call those tools yourself. You describe the look and the run applies it, the same way any Format run does, with one sandbox, one agent turn and one receipt.
| Stays | Changes |
|---|---|
| Motion | Palette (the example asks for a warm film palette) |
| Timing and cuts | Lighting (soft daylight in the example) |
| Your original audio | Product colors, taken from the reference you attach |
How it differs from recreate
sume-recreate rebuilds a reference's structure with a new presenter and everything generated from scratch. sume-restyle starts from your own clip and holds its action fixed. Choose recreate to borrow a format, restyle to re-skin a take you like.
Call it
POST /v1/formats/sume/sume-restyle/runs needs formats:write and an Idempotency-Key. Attach the packshot or look reference (up to 30 images) in attachments and describe the target look in instruction. Read the Format first with GET /v1/formats/sume/sume-restyle for its description and io profile.
Set a cap for the run. Restyling goes chunk by chunk, so a longer clip means more generation, and the effective cap is the only ceiling on it.
Check the result against the source
A restyle should match the original in motion, timing and cut points. Compare the output with your source at the cut points and at the audio, since the brief promises your original audio. If the output drifts, send the failed run's id as previous_run_id with a narrower instruction, so the next turn continues the same thread instead of starting over.
Pair it with trim and captions
If only part of a take should be restyled, cut it first with the video-trim API, which makes a new MP4 of a range up to 900 seconds from a source of up to 1800 seconds at $0.02 per job. After the restyle, burn captions with the video-captions API. Both are separate jobs, polled on /v1/jobs/{id}/status.
Whatever you chain, keep the numbers in your own config. The cap for the restyle run, the trim range and the caption job are three separate limits, and none of them knows about the others. A pipeline that has all three should fail at the first step that breaks and never go on to pay for the next one.
Sources
Related posts
More in Formats
- Season outline as nested input: one key, 64-key limit, 2 MiB
Put a whole season outline under one key of a Format run's input. Only top-level keys count toward 64, and the body can be up to 2 MiB, so 40 episodes fit.
- Shorts series sequential playback: fade out or hard cut each episode?
YouTube plays Shorts series episodes in order. Pick a hard cut or a short fade at the end of each one, with the Timeline 1.0 fade limits.
- Bulk queue item error: format_run_canceled vs format_run_failed
In a Sume bulk queue, a failed child reads format_run_failed and a canceled child reads format_run_canceled. The real reason is in the child run receipt.
- TikTok 60-minute uploads vs Sume's 30-minute source limit
Sociality.io reports TikTok uploads up to 60 minutes. Sume trims and Timelines stop at 1800 seconds (30 minutes). What fits, and the cut points for the rest.
Written by Sume