Final Cut Pro Edit Detection: sample stills to see the cuts

Edit Detection in Final Cut Pro 12.3 analyzes a rendered video to reveal shot changes and split it into clips. Sume has no shot-change detector on these routes: video inspect states it returns probe facts, stills and optional transcript, not typed scenes. What it does offer is POST /v1/video-frames with fps, so you can look at evenly spaced stills and find the cuts yourself.

Apple's claim is from its release notes; Sume's from Video frames and Video inspect, read 2026-10-01.

What does Edit Detection do?

The 12.3 notes say: analyze any rendered video with Edit Detection to reveal its shot changes and automatically split it into separate clips with just a click. It is an editor feature that outputs clips on your timeline.

What does Sume give me instead?

POST /v1/video-frames takes one media.sume.com clip plus exactly one of at[] or fps and returns durable image artifacts. It is unbilled. The docs say 0 < fps ≤ 2, expanded to mid-bin samples and capped at 24 frames; at[] takes 1 to 24 values.

Video inspect does the same with frames: { fps: n }, with the same limits, and adds a probe. Its docs say semantic scene questions are separate tools that appear only when listed in tools_list.

curl -X POST https://api.sume.com/v1/video-frames \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: cut-scan-001" \
  -d '{
    "video_url": "https://media.sume.com/artifacts/artf_demo/ad.mp4",
    "fps": 1,
    "max_edge": 480
  }'

What are the limits?

Frame budgets decide how finely you can scan a clip.

How do I find a cut from the stills?

Each frame carries t, url, width and height. At fps: 1 on a 20-second ad you get 20 stills; where two neighbors show different scenes, the cut is in that second. Re-run with at[] around that second to narrow it, then trim with a separate step, as in cut a product video into feed cutdowns.

Is this the same as automatic splitting?

No. Sume returns stills, and you or your agent decide where the boundaries are. Nothing here splits a clip for you. For stills basics see extract video frames.