Talking head video API: avatar vs lip sync vs motion control
Pick a Sume talking-head route by input: a script and a ready avatar, your own audio and a still, or a driving video. Limits and per-second prices.
Pick a Sume talking-head route by what you already have. A ready avatar and a script go to Avatar 1.0 (POST /v1/avatar-1.0/talking-video); Sume-hosted audio and a still go to lip sync (POST /v1/veed/fabric-1.0, or POST /v1/minimax/h3-max/lip-sync for 5–14.8 seconds of audio); a driving video goes to Kling 3.0 Motion Control (POST /v1/kling/3.0/motion-control).
Facts come from the Generate avatar video and Models overview docs and the request schemas in the Sume API reference, read on 2026-09-27; prices are read from Sume's pricing code. Each route has its own guide: talking avatar video, lip sync, and motion control.
Which route fits which input?
Start from what you already have, not the model name. Avatar 1.0 needs a ready avatar; lip sync and motion control take a public still in image_url or a ready avatar in avatar_id / avatar_handle. A plain video model is not a fourth option: Sume's models guide says video models do not lip-sync to generated TTS or to a later voice-over.
| Route | You bring | What drives the face | Length per job | Price |
|---|---|---|---|---|
| Avatar 1.0 talking video | A ready avatar and the words | A script, or scene-by-scene video_inputs | 4–60 s, estimated from the script | $0.184/s standard, $0.245/s plus, $0.55/s max (no product image) |
| Lip sync: VEED Fabric 1.0 or MiniMax H3 Max | Audio on the Sume media host, plus a still or an avatar | The recording in audio_url | Fabric 1–300 s of audio; H3 Max 5–14.8 s | Fabric $0.1875 per audio second (720p); H3 Max in the lip sync guide |
| Kling 3.0 Motion Control | A driving video, plus a still or an avatar | The motion in motion_video_url | Up to 30 s of driving video; the output follows it | $0.1575 per output second |
When should I use the Avatar 1.0 talking video?
When you have words but no recording. The route turns a ready avatar into a script-driven talking video: send its avatar_handle and exactly one of script or video_inputs, and Sume accepts the job when it estimates the video at 4–60 seconds. quality, aspect_ratio, and the optional product and scene inputs are covered in the talking avatar video guide.
- To check first-frame stills before paying for the full render, create an avatar video preview, then call
generate-videoon its id. - For captions, burn them onto the finished video with
POST /v1/video-captions.
When is lip sync the better choice?
When the voice already exists as a recording, such as a TTS line. Lip sync animates the face to that audio, with no script involved. Fabric takes up to 300 seconds of audio per job, and H3 Max Lip Sync is the explicit alternative inside a 5–14.8 second window. The lip sync guide covers both request bodies and their prices.
What does motion control do differently?
It animates a still with the motion of a driving video you supply, so the performance comes from footage instead of a script or an audio file. Send motion_video_url (public HTTPS, at most 30 seconds), its length in duration_seconds, and one visual source. The output length follows the driving video.
keep_original_sound keeps the driving video's audio track unless you set it to false, and prompt only steers appearance details. The motion control guide covers the other fields.
What if my video is longer or has two speakers?
Then the per-job limits decide the plan:
- Past 60 seconds of script, Avatar 1.0 needs several jobs: the docs say to split a longer script, and How to make an AI avatar video longer than 60 seconds joins the parts.
- Up to 300 seconds of finished audio fits one Fabric job, and up to 30 seconds of driving video fits one motion-control job.
- Two speakers need two jobs, because current Avatar 1.0 execution supports one resolved avatar per final video. AI avatar conversation video cuts the turns together.
- Silent pauses can stay inside one Avatar 1.0 job: a
video_inputsscene withvoice.type: "silence"is a non-speaking beat.
How do the prices compare?
Every route bills per second, plus a 5.5% agent fee by default, but the seconds are counted differently. Lip sync counts audio seconds and motion control counts driving-video seconds, each rounded up. Avatar 1.0 bills at a rate set by its quality tier, and the rate rises with a product image; the product rates are on API pricing.
What do the routes have in common?
They share the job contract, so switching routes changes the body, not your polling code:
- Each create runs as a job. Send an
Idempotency-Key, store the job id, poll the status, and read Sume-hosted artifacts from the result. - On hosted MCP, Avatar 1.0 is
avatar-videos_create, Fabric isavatar-image-to-video_create, and motion control iskling-motion-control_create. The docs' hosted tool inventory lists no H3 Max Lip Sync tool. - To put an avatar's face onto an existing video instead, Face swap is in Beta.
Sources
Related posts
More in Sume Avatar 1.0
- Introducing Sume Avatar 1.0
Sume Avatar 1.0 is a multi-agent orchestration system as a single avatar model.
- Avatar Face Swap API (Beta): apply an avatar face to a video
Avatar Face Swap 1.0 is a Beta Sume endpoint that applies a ready avatar's face to a short public source video. Required fields, limits, and polling.
- Avatar video previews: approve the first frame before rendering
Create an avatar video preview to get first-frame stills, regenerate them if needed, then call generate-video on the preview id to render the final video.
- Multi-scene avatar video API: build one video from ordered scenes
Send ordered video_inputs instead of one script to compose spoken and silent scenes into one 4-60 second avatar video. Scene fields, rules, and limits.
Written by Sume