Swap four people in one video: H3 Max Recast bills per second

h3-max-recast accepts one source video and one photo per person, from 1 to 4 people, and the output keeps the source length of 5 to 30 seconds. On Sume the price depends on seconds and resolution only: $0.375 per second at 768p, $0.5625 at 1080p. The docs list no per-person charge, and the estimator that reserves your balance takes only duration and resolution.

Price by source length

The reserved duration is the source clip length, rounded up to a whole second.

Photo rules

• One photo per person you want in the output, 1 to 4 photos.
• The prompt is optional.
• Resolution is 768p or 1080p; the Sume default is 768p.
• Video references are accepted; audio references are not.

Request

Recast runs through the Video Router wire with video_url and reference_image_urls.

curl -X POST https://api.sume.com/v1/video-router/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: recast-two-people-001" \
  -d '{
    "model": "h3-max-recast",
    "video_url": "https://example.com/source.mp4",
    "reference_image_urls": [
      "https://example.com/person-a.jpg",
      "https://example.com/person-b.jpg"
    ],
    "resolution": "768p",
    "duration": 12,
    "mode": "async"
  }'

Planning a run

A 20-second source at 768p is $7.50; at 1080p it is $11.25. Because recast bills the source length, trim the source to the shot you need before you submit. Video trim is a $0.02 job, which is cheaper than paying for seconds you will cut.

Sume reserves the full price at submit, so the number in the table is also the amount that must be free in the balance when you send the request. If it is not, the call fails with 402 insufficient_credits before any provider work starts.