Demand Gen video crop and trim tool vs an API crop and trim

Google's Demand Gen tool crops a video to horizontal (16:9), square (1:1) or vertical (9:16) by moving a crop area, and trims it to between 5 and 300 seconds with sliders, inside Google Ads. To do the same outside the console, Sume takes two calls: a crop op on video filter, then video trim.

Google facts are from its Help page; Sume facts from the two docs pages. Read 2026-09-30. Ad specs themselves are in Demand Gen video ad specs.

What does the console tool do?

You open the pencil icon on an uploaded video in the ad builder, pick an aspect ratio, move the crop area, drag the trim sliders, preview, and choose Save as new video. Uploaded videos must be at least 5 seconds long.

How does each step map to Sume?

How do I write the crop?

Crop bounds are fractions, not pixels. width and height run from 0.05 to 1, and x+width and y+height must not exceed 1. For a 16:9 source cropped to 9:16, take the full height and a width of 9/16 divided by 16/9, about 0.316, centered at x of 0.342. The compiler even-rounds the result for yuv420p, so the pixel size can differ by one from your arithmetic.

curl -X POST https://api.sume.com/v1/video-filter \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: demand-gen-crop-001" \
  -d '{
    "video_url": "https://media.sume.com/artifacts/artf_demo/talk.mp4",
    "ops": [{ "op": "crop", "x": 0.342, "y": 0, "width": 0.316, "height": 1 }]
  }'

Which order should I run them in?

Trim first if the source is over 300 seconds, because the filter's source cap is 300 s while trim accepts sources up to 1800 s. Feed the new video_url from the first job's result into the second call, and poll GET /v1/jobs/:id/result for each. Neither call previews; check the output before uploading it to Google Ads.