Avatar face swap Beta: check the 4 to 15 second clip before you submit
Sume's Avatar Face Swap Beta needs a public HTTPS source clip of about 4 to 15 seconds with usable audio and a required quality tier.
What the Beta accepts
Avatar Face Swap 1.0 applies a ready avatar face to a public HTTPS source video. The required fields are avatar_handle, video_url and quality, and the Beta worker targets source clips of about 4 to 15 seconds with usable audio.
It is a Beta model-run endpoint, not the old consumer /face-swap route. Use it when you already have an Avatar 1.0 identity and a short clip, and you do not need a script-driven talking video; for that, use the avatar-video route.
Request rules
| Item | Rule |
|---|---|
| video_url | fetchable public HTTPS video; no localhost, private, signed or provider-task URLs |
| quality | required: standard, plus or max; no default |
| Unsupported | prompts, transcripts, durations, aspect ratio, avatar ids in the body, provider fields |
curl -sS -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-holiday-001" \
-d '{
"avatar_handle": "holiday_host",
"video_url": "https://cdn.example.com/clips/greeting-12s.mp4",
"quality": "standard"
}'Check the clip first
A free pre-check saves a failed job. Probe the duration and audio of your clip before you submit, and trim it to the window if it is long. Sume has Video trim for cutting to a start and end and Audio detach to check the audio track.
Read the result safely
Poll /v1/jobs/{id}/status, fetch /result when completed, and prefer resource_status for readiness and job_status for polling. Communication modes follow the other generation submits: async, sync or subscribe with wait_timeout_seconds, and webhook with a public HTTPS URL. Use the face of a person only with their consent, and disclose edited faces where the platform's rules require it.
Sources
Related posts
More in Sume Avatar 1.0
- Does Sume face swap keep the original voice, outfit and scene?
Sume's Face Swap (Beta) keeps the source clip's audio, camera and timing but replaces the whole person, not just the face. What stays, what changes, cost.
- Face swap API webhook: get job.completed instead of polling
Sume's face swap (Beta) takes mode webhook with a public HTTPS webhook_url, then sends one signed terminal event: job.completed, job.failed or job.canceled.
- Face swap for UGC ads: source clip rules (Sume beta)
Sume's beta Avatar Face Swap applies a ready avatar face to a public 4-15 second source video with audio. Required fields, quality tiers and what it rejects.
- HeyGen lipsync captions are always on; Sume's are opt-in
HeyGen deprecated enable_caption on translation and lipsync and now always returns SRT and VTT. Sume avatar videos burn captions only when you ask for them.
Written by Sume