Shorts conversational editing vs API rough cuts with Trim

Conversational editing in the YouTube app is a good fit for one-off edits by hand. For rough cuts you repeat across many clips, Sume's video trim takes one clip and a range, returns a new MP4 holding only that range, and costs $0.02 per job. Each cut is a recorded API call with an id.

What YouTube announced

Reports on Made on YouTube 2026 from Sep 23, 2026 list conversational AI editing in Shorts, expanded likeness detection, real-time live auto-dubbing and Shorts series on TV. They do not give limits or availability, so check YouTube for specifics.

What Trim does

POST /v1/video-trim takes one media.sume.com clip, a start in seconds, and exactly one of end or duration. The source is untouched and a new artifact is returned. The video must already be in your workspace; there is no open-internet fetch, so import first with POST /v1/media-imports. Idempotency-Key is required.

curl -X POST https://api.sume.com/v1/video-trim \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: video-trim-001" \
  -d '{
    "video_url": "https://media.sume.com/artifacts/artf_demo/talk.mp4",
    "start": 2,
    "duration": 8
  }'

Options and limits

Defaults are precision: "exact", a frame-accurate re-encode, and audio: "keep". precision: "keyframe" is a stream copy whose cut may start a GOP early; re-base against actual_start_seconds. An output conform of width, height and fps works with exact only. The source may be up to 1800 seconds and the output 0.2 to 900 seconds. An end past the source clamps, and the result warns trim_clamped_to_source.

Where an avatar clip fits

If the clip is a talking presenter, make it with POST /v1/avatar-1.0/talking-video (scripts of 4 to 60 seconds), then trim or assemble it. Timeline 1.0 joins clips over one audio spine; captions can be burned last. Poll GET /v1/jobs/{id}/status for each job and read the new MP4 from the result.