Face swap on a short clip with Sume Avatar 1.0: a 4 to 15 s Beta

It can, with limits. Avatar Face Swap 1.0 is a Beta endpoint that applies a ready avatar's face onto a public HTTPS source video. The docs say the Beta worker currently plans for source videos of about 4 to 15 seconds with usable audio, so it fits a single hook shot, not a full ad.

What the call needs

• avatar_handle of a ready Avatar 1.0 identity.
• video_url: a fetchable public HTTPS URL. Localhost, private networks, non-HTTP, signed or private URLs, and provider task URLs are rejected.
• quality: standard, plus or max. It is required in Beta, with no default.
• Nothing else: prompts, transcripts, duration knobs, aspect ratio, avatar ids in the body and provider fields are unsupported.

Cost

The pricing code reserves the maximum source length, 15 seconds, at the tier rate. At the no-product Avatar Video rates that is 15 x $0.184 = $2.76 on standard, $3.68 on plus and $8.25 on max. Treat these as the reservation ceiling and confirm live in GET /v1/catalog, since the docs page does not print a face-swap price.

Face swap or talking video?

Face swap does not read a script. If you want the presenter to say new words, use the avatar video route. Only use a face you have rights to; the avatar is the identity you control.

Request

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: face-swap-1" \
  -d '{"avatar_handle":"your_avatar","video_url":"https://example.com/hook.mp4","quality":"standard"}'