How to shorten a video: trim, cut, or remove pauses

To shorten a video, cut out the time you don't need. If the extra time is at the start or end, trim it off. If it's in the middle, cut those sections out and join what's left. If a talk drags all the way through, cut the pauses between sentences.

The Sume facts below come from the Video trim, Timeline 1.0, Audio detach, Video inspect and Video filter docs and the Sume API reference, and the FFmpeg facts from the FFmpeg filters documentation, all read on 2026-09-28. Anything described as current behavior is read from Sume's code. If you want a smaller file rather than a shorter video, see how to reduce video file size.

Which way should I shorten my video?

Pick by where the extra time is. With Sume, all three work on files already in your workspace on media.sume.com, such as an earlier Sume job's output; there is no public upload route for a file on your computer (which URLs each endpoint accepts).

How do I trim the start or end of a video?

Send the range you keep. start is the new in-point in seconds, and exactly one of end or duration closes the range. The result is a new MP4 holding only [start, end), and the source is untouched. The default exact precision re-encodes and cuts on the exact frame; keyframe copies without re-encoding, but the cut may start up to a GOP early. This request drops the first 4 seconds and everything after 1:32:

curl -X POST https://api.sume.com/v1/video-trim \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: promo-shorten-001" \
  -d '{
    "video_url": "https://media.sume.com/artifacts/artf_demo/promo.mp4",
    "start": 4,
    "end": 92
  }'

How do I cut sections out of the middle?

A trim keeps one range per job, so it can't close a gap. Render the kept ranges back to back in one Timeline 1.0 job instead. Each range becomes a video[] slot that reads the source from its source_in, plus an audio.parts[] slice of the video's detached sound: in current code, a render drops each slot's own audio and plays only its audio spine. Remove part of a video by API walks through the request.

How do I cut the pauses?

Transcribe the video with POST /v1/video-inspect and transcribe: true. Each entry in transcript.words[] carries start and end in seconds from the start of the video, so a long gap between one word's end and the next word's start is a pause. Keep the speech ranges and render them as above. How to remove silence from a video turns word timings into cuts.

Can I speed the video up instead?

Not with Sume's media tools. The video filter's allowlist leaves out setpts, and a Timeline slot has no speed field. On your own computer, FFmpeg changes speed with filters: its docs call setpts=0.5*PTS a fast motion effect for video, and atempo sets audio tempo from 0.5 to 100.

What are the limits?

• A trim reads sources up to 1,800 seconds and keeps 0.2 to 900 seconds per job. To keep more than 900 seconds, use a Timeline render, whose output runs up to 1,800 seconds.
• In current code, trim, inspect, detach and the render all refuse a source file over 300 MiB with source_too_large.
• An end past the source clamps to the source, and the result warns trim_clamped_to_source.
• A trim is $0.02 per job, a detach $0.01 per job and a render $0.10 per output minute, plus a 5.5% agent fee by default. A transcript is billed per audio minute; the inspect probe is unbilled.