InfiniteTalk
Long-form talking video from one image and up to 10 minutes of audio at 480p or 720p.
Lip sync · Talking avatar · REST API
Lip Sync API
Send one portrait and one voice track and get back a lip-synced talking video, up to 10 minutes long with InfiniteTalk or up to 2 minutes with Seedance 2.5 Talking Avatar. One API key, pay-as-you-go credits and an exact quote before you render.
cURL · InfiniteTalk
curl --request POST \
https://www.makeinfluencer.ai/api/v1/models/infinitetalk/generations \
--header "Authorization: Bearer $MAKEINFLUENCER_API_KEY" \
--header "Idempotency-Key: infinitetalk-001" \
--header "Content-Type: application/json" \
--data '{
"input": {
"startFrameUrl": "https://example.com/portrait.png",
"audioUrl": "https://example.com/voiceover.mp3",
"resolution": "480p",
"durationSeconds": 30
}
}'How It Works
Get Started
Create an API key
In Dashboard → Developer API → Keys, create a revocable mi_ key and send it as a Bearer token.
Upload and preview
Upload the portrait and audio with POST /assets/presign, then quote the request with preview-credits. Previews never charge.
Generate and poll
POST to /models/{modelId}/generations, receive HTTP 202 with a generation ID, then poll the video status endpoint until terminal is true.
Lip sync API pricing
Talking avatar models, quoted by audio length
Both models turn one portrait and one audio track into a lip-synced video. Cost scales with the length of the audio and the output resolution, so each row quotes a real request.
| Model ID | Model | Quoted request | Max audio | Typical time | Credits |
|---|---|---|---|---|---|
| infinitetalk | InfiniteTalk | 30 s of audio, 480p | Up to 10 min | varies | 171,429 |
| infinitetalk | InfiniteTalk | 30 s of audio, 720p | Up to 10 min | varies | 216,542 |
| infinitetalk | InfiniteTalk | 120 s of audio, 480p | Up to 10 min | varies | 685,715 |
| seedance25_talking_avatar | ByteDance Seedance 2.5 Talking Avatar | 30 s of audio, 480p | Up to 2 min | ~2 min | 330,828 |
| seedance25_talking_avatar | ByteDance Seedance 2.5 Talking Avatar | 30 s of audio, 720p | Up to 2 min | ~2 min | 661,655 |
Quotes are calculated from the live model pricing when this page is built. The server measures the uploaded audio for the final charge. POST /models/{modelId}/preview-credits returns the exact cost of your own payload without charging.
Quickstart
A talking avatar from a photo and a voice track
Upload a portrait and an audio file, pass both URLs, and the POST returns HTTP 202 with a generation ID. Poll the video status endpoint until terminal is true and download the MP4.
POST
https://www.makeinfluencer.ai/api/v1/models/infinitetalk/generationscURL
curl --request POST \
https://www.makeinfluencer.ai/api/v1/models/infinitetalk/generations \
--header "Authorization: Bearer $MAKEINFLUENCER_API_KEY" \
--header "Idempotency-Key: infinitetalk-001" \
--header "Content-Type: application/json" \
--data '{
"input": {
"startFrameUrl": "https://example.com/portrait.png",
"audioUrl": "https://example.com/voiceover.mp3",
"resolution": "480p",
"durationSeconds": 30
}
}'POST · preview-credits · no charge
curl --request POST \
https://www.makeinfluencer.ai/api/v1/models/infinitetalk/preview-credits \
--header "Authorization: Bearer $MAKEINFLUENCER_API_KEY" \
--header "Content-Type: application/json" \
--data '{
"input": {
"startFrameUrl": "https://example.com/portrait.png",
"audioUrl": "https://example.com/voiceover.mp3",
"resolution": "480p",
"durationSeconds": 30
}
}'GET · poll until terminal
curl \
https://www.makeinfluencer.ai/api/v1/generations/video/GENERATION_ID \
--header "Authorization: Bearer $MAKEINFLUENCER_API_KEY"Features
What you get with the lip sync API
01
One image, one audio track
startFrameUrl and audioUrl are the only required inputs. No video footage, no face tracking setup and no model training.
02
Any voice
Use a recording, a text-to-speech voice or a dubbed track in another language. The mouth follows the audio you send.
03
Up to 10 minutes per video
InfiniteTalk handles long narration, courses and podcasts up to 10 minutes, and Seedance 2.5 Talking Avatar handles up to 2 minutes with richer gestures.
04
Signed uploads
POST /assets/presign returns a signed URL for images and audio, so you never have to host files yourself.
05
No separate API plan
REST, MCP and the dashboard draw from the same credit balance at the same model prices, with no monthly API fee.
06
Also in Claude and Cursor
The same models are available through the MCP connector, so an agent can generate talking videos on your account.
Price check
What InfiniteTalk lists for on fal.ai
Public list prices for comparable requests elsewhere, verified 29 September 2026, next to our credit quote for the same request. Our credits come from the balance you already use in the app, and preview-credits returns the exact figure before you spend anything.
| Model | Provider | Their list price | Their example | Our model ID | Our credits, same request |
|---|---|---|---|---|---|
| InfiniteTalk | fal.ai | $0.20/s at 480p, $0.40/s at 720p | $6.00 for 30 s at 480p | infinitetalk | 171,429 (30 s of audio, 480p) |
Third-party prices are copied from each provider's public model page on 29 September 2026 and can change at any time. Sources: https://fal.ai/models/fal-ai/infinitalk.
Choosing a lip sync model
InfiniteTalk or Seedance 2.5 Talking Avatar?
Use InfiniteTalk for long narration, courses and podcasts up to 10 minutes. Use Seedance 2.5 Talking Avatar for short ads and social clips where gestures and expressions matter most.
Seedance 2.5 Talking Avatar
Natural lip sync with gestures and expressions for up to 2 minutes of audio.
Any voice source
Record it, or generate it with any text-to-speech tool, then pass the audio URL.
Practical workflow
- 1
Create a key
Dashboard → Developer API → Keys. Keys start with mi_.
- 2
Upload portrait and audio
POST /assets/presign for each file, then pass the URLs.
- 3
Preview and generate
Quote the body with preview-credits, then POST it to generations.
- 4
Poll and store
Poll GET /generations/video/{id} and save the url when ready is true.
InfiniteTalk is an open-source talking-video model built on Alibaba's Wan video model, and Seedance is a ByteDance model family. MakeInfluencer.ai is an independent API provider and is not affiliated with or endorsed by their creators.
Trusted by over 300,000+ creators
FAQ
Frequently Asked Questions
Add talking avatars to your product today
Create an API key, upload a portrait and a voice track and render your first talking video in a few minutes.
Start creating in minutes · Pay as you go