MiniMax H3 Lip Sync
MiniMax H3 Lip Sync turns one photo into a talking video from your audio. The mouth follows the audio, and the output is as long as the audio. It renders at 480p, 768p, 1080p or 2K.
lipsync · model h3-max-lipsync · catalog model-catalog@2026-10-06.v5 · pricing customer-pricing@2026-11-05.v15.Endpoint
POST https://api.roasos.com/v1/lipsync/h3-max-lipsyncSend your API key as a bearer token. The key needs the scope api:write. Every paid request also needs an Idempotency-Key header.
The envelope fields quote_only, quote_id, webhook_url and wait_timeout_seconds behave the same on each model endpoint. Refer to Jobs and Quotes and credits.
Good at
A UGC-style talking head from a product photo of a person, a saved avatar or a creator still, paired with a voiceover. Send image_url (or avatar_id) and audio_url. The audio must be at least 5 seconds long, and the product accepts at most 60 seconds. The image aspect ratio must be between 0.4 and 2.5; the output uses the nearest supported ratio. resolution sets the tier and the price. Clips longer than 15 seconds cost 1.2 times the per-second rate. It has no prompt.
Pricing
The unit is one second. Pricing key: h3-max-lipsync.
| Request | Credits |
|---|---|
10s | 10 |
The credits come from the roasOS pricing table. No other field changes the price. To get the exact price of your request, send it with quote_only: true first.
Modes
image-audio-to-video
Provider endpoints: minimax/h3-max/lip-sync/image-to-video.
Request fields
| Field | Type | Required | Default | Rule |
|---|---|---|---|---|
audio_url | string | yes | — | https URL; ≤ 4,096 characters |
avatar_id | string | no | — | UUID |
image_url | string | no | — | https URL; ≤ 4,096 characters |
quote_id | string | no | — | UUID |
quote_only | boolean | no | false | — |
resolution | enum | no | 768p | 480p, 768p, 1080p, 2k |
wait_timeout_seconds | integer | no | 0 | 0–60 |
webhook_url | string | no | — | https URL; ≤ 4,096 characters |
The schema is strict. A field that is not in this table gets 400 unsupported_option. The field names are the roasOS role names. One role has one name in every family.
Media
This endpoint accepts these media roles:
avatar_id— identity; the saved character reference sheet, with a derived portrait only where the model requires a single face or source frameimage_url— first frame (or the single source image)audio_url— the driving audio track
| Role | Items | Maximum size | Content types | Duration |
|---|---|---|---|---|
audio_url | ≤ 1 | 15 MB | audio/mp3, audio/mpeg, audio/wav, audio/wave, audio/x-wav | 0–60 s |
image_url | ≤ 1 | 10 MB | image/jpeg, image/png, image/webp | — |
Every other role is refused: elements, product_id, product_images, references, reference_images, end_image_url, scene_id, scene_image, video_urls, video_url, audio_urls, voice_id. A refused role gets 400 unsupported_option.
https://cdn.roasos.com that your workspace owns. Any other URL gets 400 reference_media_not_workspace_owned. Refer to Reference media and uploads to learn how to put a file there.Prompt
None. The endpoint has no prompt. The audio drives the performance, so record or generate the voiceover with the delivery you want.
Not exposed
These catalog entries describe omitted inputs, pinned values and source constraints. They do not imply that the provider accepts every listed field.
enable_transcription— pinned toprovider default (off); lips follow the audio itself.enable_safety_checker— pinned toprovider default (on); never sent.seed— not accepted by this public endpoint.
Result
The result kind is lipsync. Always present: url.
{
"data": {
"id": "gen_…",
"kind": "lipsync",
"model": "h3-max-lipsync",
"status": "completed",
"result": {
"url": "https://cdn.roasos.com/workspaces/ws_01/result"
}
}
}The files are immutable objects on https://cdn.roasos.com. A result URL is a valid media input for your next request.
Example
The three steps are always the same: quote, run, poll.
# 1. Quote. Nothing is spent.
curl -X POST https://api.roasos.com/v1/lipsync/h3-max-lipsync \
-H "Authorization: Bearer roasos_sk_live_..." \
-H "Idempotency-Key: 0f1c9e2a-h3-max-lipsync-0001" \
-H "Content-Type: application/json" \
-d '{"quote_only":true,"audio_url":"https://cdn.roasos.com/workspaces/ws_01/uploads/audio_url.mp3","image_url":"https://cdn.roasos.com/workspaces/ws_01/uploads/image_url.png"}'
# 200 { "data": { "kind": "quote", "quote_id": "q_7Hk...", "credits": 10, "expires_at": "...",
# "confirmation": { "model": "h3-max-lipsync", "example": "10s" } } }
# 2. Run. Same body, same Idempotency-Key, quote_id in place of quote_only.
curl -X POST https://api.roasos.com/v1/lipsync/h3-max-lipsync \
-H "Authorization: Bearer roasos_sk_live_..." \
-H "Idempotency-Key: 0f1c9e2a-h3-max-lipsync-0001" \
-H "Content-Type: application/json" \
-d '{"quote_id":"q_7Hk...","audio_url":"https://cdn.roasos.com/workspaces/ws_01/uploads/audio_url.mp3","image_url":"https://cdn.roasos.com/workspaces/ws_01/uploads/image_url.png"}'
# 201 { "data": { "id": "gen_...", "status": "queued", "credits_reserved": 10 } }
# 3. Poll, then read the result.
curl https://api.roasos.com/v1/jobs/gen_.../status -H "Authorization: Bearer roasos_sk_live_..."
curl https://api.roasos.com/v1/jobs/gen_.../result -H "Authorization: Bearer roasos_sk_live_..."The deprecated path
POST /v1/generations still reaches this model for one minor version. It is deprecated. This is how it maps:
The endpoint is POST /v1/lipsync/h3-max-lipsync with image_url (or avatar_id), audio_url and resolution (480p|768p|1080p|2k).
The deprecated alias does not carry this model: POST /v1/generations { "type": "lipsync" } returns 503 lipsync_public_authority_unavailable for every lipsync model except ltx-lipsync.
See also
llms.txtfor this model — the same facts as one plain-text file for an agent.GET https://api.roasos.com/v1/catalog/h3-max-lipsync— the machine-readable form: enums, defaults, limits and prices. Refer to the API reference.- Jobs — how to poll a job and read its result.
- Provider endpoints:
minimax/h3-max/lip-sync/image-to-video.
Seed Audio 1.0
Text-to-audio from ByteDance: speech with 20 preset voices, or a read that matches up to three reference clips or one reference image. It is best at scene audio, multi-voice reads, and expressive delivery.
HeyGen Photo Avatar 4
HeyGen Photo Avatar 4 animates a photo with a clear face to an audio track and controls the output frame directly: a resolution preset, an aspect ratio and a talking style.