InfiniteTalk API

by MeiGen-AI

Talking video up to 10 minutes from a portrait or clip at 480p or 720p.

meigen-ai/infinitetalk-v1/talking-avatar
  • Up to 720p
  • AI Avatar
  • Lip Sync

About InfiniteTalk

InfiniteTalk is an open-source model from MeiGen-AI, released in August 2025 under the Apache-2.0 license. It turns one photo and a voice recording into a video of that person speaking, at 480p or 720p, for recordings from 3 seconds to 10 minutes. In a group photo, a mask marks which person should talk.

Call it with the model name meigen-ai/infinitetalk-v1/talking-avatar on POST /jobs/createTask, from $0.039.

Introduction

InfiniteTalk takes image_url for a photo and audio_url for the spoken recording. It returns a video with synchronized speech and subtle natural movement. MeiGen-AI provides the model as open source under Apache-2.0.

Key features

The recording can run from 3 seconds to 10 minutes. When several people appear in the photo, use mask_url to mark the intended speaker in white.

In our tests, the rooftop astronomer delivered an approximately nine-second explanation of Saturn beside a telescope. The 480p clip followed the audio with small natural movements. That sample used a prompt and did not include a speaker mask or seed.

Request parameters

ParameterRequiredDescription
audio_urlYesThe voice or music track to use.
image_urlYesThe picture to start from.
mask_urlNoPaint over the area that should change.
promptNoDescribe what you want to make.
resolutionNoChoose an available output size. 480p · 720p · Default: 480p
seedNoUse the same seed and settings for more controlled comparisons. ≥ 0 · ≤ 2147483647

Pricing

Prices are in USD and charged to your PopNow balance. When a task is accepted, we reserve the maximum quoted amount from your API balance. On completion, we charge for actual usage, never more than the hold, and release the remainder. If the task fails, the hold is released.

OptionUSD per second
resolution 480p$0.039
resolution 720p$0.068

Send the same request body to POST /jobs/quote to see the estimated and maximum USD charge before running. A quote does not run the task or charge your balance.

Use cases

Animate a recorded astronomy explanation using a portrait beside a telescope. Select a speaker in a group image through a white mask. Compare a single change in direction or resolution with the same photo, recording, and random source.

Prompt guide & pro tips

Describe the scene and intended movement in the optional prompt; provide the spoken material through audio_url. For a group photo, mark the speaker in white in the mask rather than relying on a scene description to identify them.

Reuse seed to keep the same random source when comparing one change. A repeated seed does not guarantee the same video frame for frame. Use photos and voices with the necessary rights, only for yourself or for people who have agreed to this use.

API quick start

Call the InfiniteTalk API in three steps

  1. Create an API key. Sign in to PopNow and create a key in the API console. Keys start with sk-popnow-. API calls are billed in USD from your PopNow balance.
  2. Send the request. POST /jobs/createTask with model meigen-ai/infinitetalk-v1/talking-avatar and an input object that follows the parameter table. The response returns a task ID straight away.
  3. Get the result. Poll GET /jobs/recordInfo with the task ID, or pass a callBackUrl and we post the result to it. Download the output links when the task succeeds.
# 1. Create the task. Keep the Idempotency-Key if you retry after a network error.
export IDEMPOTENCY_KEY="$(uuidgen)"
curl -X POST https://api.popnow.ai/api/v1/jobs/createTask \
  -H "Authorization: Bearer $POPNOW_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $IDEMPOTENCY_KEY" \
  -d '{
  "model": "meigen-ai/infinitetalk-v1/talking-avatar",
  "input": {
    "image_url": "https://media.popnow.ai/influencer/examples/marco.webp",
    "audio_url": "https://media.popnow.ai/shortfilm/voices/Aoede-46e4c427.mp3",
    "resolution": "480p"
  }
}'

# 2. Copy data.taskId from the response, then poll until state is succeeded, failed, canceled or expired.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.popnow.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
  -H "Authorization: Bearer $POPNOW_API_KEY"

Replace the prompt and any links with your own. The API tab has the response, polling and callback reference.

InfiniteTalk API questions

How much does the InfiniteTalk API cost?

InfiniteTalk starts at $0.039 per second with AI Avatar, its lowest-priced task. Prices by option: 480p $0.039 per second, 720p $0.068 per second. API usage is charged in USD to your PopNow balance. When a task is accepted, we reserve the maximum quoted amount from your API balance. On completion, we charge for actual usage, never more than the hold, and release the remainder. If the task fails, the hold is released.

What is the model name for InfiniteTalk in the API?

Put meigen-ai/infinitetalk-v1/talking-avatar for AI Avatar; meigen-ai/infinitetalk-v1/lip-sync for Lip Sync in the model field of POST /jobs/createTask.

Which InfiniteTalk endpoint fits a photo or a video?

Choose talking-avatar for a still portrait and lip-sync for existing footage. Each needs a recording; the video endpoint can regenerate expressions and posture as it matches the new speech.

Can an InfiniteTalk recording last ten minutes?

Yes, the accepted audio range is 3–600 seconds. Output duration follows the recording, and a longer source video is shortened to fit; the florist example used 6.1 seconds of speech, so it does not verify ten-minute performance.

How do I identify the speaker when several people appear?

Provide a mask image with the intended person marked white. The optional mask_url control is available for the photo and video workflows; the florist request did not use a mask, so that example does not verify group-speaker selection.

Can a fixed seed help me compare motion instructions?

Yes, a fixed seed helps isolate a single changed setting. Keep the media and resolution steady when revising a motion cue, and inspect both renders; matching seed values do not guarantee matching results.

Does the prompt supply the words or generate a voice?

The uploaded recording supplies the words and voice. InfiniteTalk does not clone voices; use prompt for visible expression, movement, or lighting, with up to 2,000 characters on the video endpoint.

What are the limits of the InfiniteTalk API?

Resolutions: 480p · 720p; Prompt length (characters): 2,000; Largest upload: 50 MB. Requests outside these limits are rejected before anything is charged.

Can I try InfiniteTalk without writing code?

Run InfiniteTalk in the Playground with your API USD balance, without an API key. Each run appears in your developer history with its matching code.

More about InfiniteTalk