Swapping a face in a 12-second clip: manual steps vs one request

With Sume you replace a face in a short clip with one request: POST /v1/models/sume/avatar-face-swap/v1.0/runs with an avatar_handle, a public HTTPS video_url and a quality. The beta is meant for source videos of roughly 4 to 15 seconds with usable audio (Face swap (Beta)).

That is a description of the shape of the work, not a verdict on doing it manually. Here is what each route asks of you.

Side by side

The request

quality has no default in this beta; leaving it out is an error. The endpoint takes no prompt, transcript, duration, aspect ratio or provider fields.

curl -X POST https://api.sume.com/v1/models/sume/avatar-face-swap/v1.0/runs \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: swap-001" \
  -d '{
    "avatar_handle": "reference_presenter",
    "video_url": "https://example.com/source-12s.mp4",
    "quality": "plus"
  }'

Cost ceiling

The beta reserves the maximum source length (15 s) at the Avatar Video no-product tier rate, so this is the most one swap can hold.

Reading the result

Poll /v1/jobs/{id}/status, then use resource_status as the primary readiness signal and job_status while polling. A completed resource exposes a video_url under media.sume.com.