Join voiceover takes into one gapless track with Timeline audio
Concatenate up to 20 Sume-hosted voice takes with Timeline audio, with no seam silence and no re-synthesis, and re-base video starts from the returned offsets.
Use POST /v1/timeline-1.0/audio with operation: "concat" and an ordered parts[] of 1 to 20 takes. Sume joins them in the sample domain, so there is no re-synthesis and no silence added at the seams. It costs $0.01 per job and returns one reusable audio file plus the offset of each part.
Request rules
| Rule | Detail |
|---|---|
| parts | 1 to 20, ordered; each has url, optional source_in and duration |
| Hosting | URLs must be this workspace's media.sume.com audio; import first |
| Idempotency-Key | required header |
| Channel layout | must match across parts, or audio_parts_channel_mismatch |
| Top-level url / ranges | not allowed on concat |
| Output | wav or mp3, up to 1,800 s |
Trim a take without re-recording
Each part can carry source_in (start inside that file, in seconds) and duration. Use them to cut a breath or a restart from the head or tail of a take. In the docs example, a part starts 0.1 s in and runs 1.8 s.
Using the offsets
The result has audio_url, duration_seconds and segments[], each with index, start and duration_seconds. Those are the concat offsets to re-base your Timeline 1.0 video[].start values against, so a scene cut lands on the line that goes with it. Pass audio_url as audio.url on a render.
When to skip the job
If the join is only needed inside one render, Timeline 1.0 accepts audio.parts[] directly, and the separate concat job is unnecessary. Mint the file only when you want to reuse it, for instance as input to another render or to avatar image-to-video.
Limits
More than 20 takes needs two concats and a third to join the results. Mixed mono and stereo parts fail, so generate every take with the same output format. Ready-made takes for a halloween-style story are in the Halloween narration post; for long scripts, see splitting scripts that exceed the duration cap.
More in Media tools
- Crop a 16:9 video to a centered 9:16 strip: the crop fractions
For a centered 9:16 crop of a 16:9 video, send crop x 0.3418, y 0, width 0.3164, height 1 to Sume video-filter. 1:1 and 4:5 values are in the table.
- Cyber Monday ad in 9:16, 1:1 and 16:9 from one clip with Timeline
Resize one offer video to vertical, square and landscape with Timeline 1.0 output width and height, fit blur or cover, for $0.30 across three renders.
- Edit Video in Sume: which models appear and why H3 Max does not
Sume's Edit Video mode lists Auto, Kling 3.0, Wan 3.0 and MiniMax H3 and needs a reference video. H3 Max and Grok Imagine are missing; the API differs.
- ElevenLabs sound effects API: 0.5-30 s, loop, and Sume's route
ElevenLabs POST /v1/sound-generation takes duration_seconds 0.5-30 and a loop flag. Sume lists no sound-effect endpoint; here is the music route and its limits.
Written by Sume