muapi Logomuapi Logo
muapi Logomuapi Logo

OmniHuman 2.0: AI Lipsync

Animate a portrait with audio using OmniHuman 2.0: realistic lip sync, prompt control, up to 300s. Try free — pay per generation, no subscription. Generate realistic talking and singing avatar video from a portrait image and audio using OmniHuman 2.0. Optional text prompt controls scene, motion and camera.

Interactieve modelbediening

Generate realistic talking and singing avatar video from a portrait image and audio using OmniHuman 2.0. Optional text prompt controls scene, motion and camera.

📝

Overview

Over dit model

OmniHuman 2.0 animates a single portrait image with an audio track into a realistic talking or singing video. Motion and lip sync are driven by the audio, and an optional text prompt of up to 300 characters controls the scene, character motion and camera movement. It accepts audio from 4 to 300 seconds, so it suits long-form presenters, explainers and music clips, with prompts in Chinese, English, Japanese, Korean, Spanish and Indonesian.

1Digital Avatars: Create presenters and spokespeople from one portrait and a voiceover.
2Education: Turn lesson audio into a talking instructor video up to five minutes long.
3Music: Make a portrait sing along to a track with natural lip sync.
4Marketing: Produce multilingual spokesperson clips by swapping the audio.
💰

Prijzen en waarde

Cost analysis

muapiapp$0.10 per second of audio

Audio duration rounded up to the next second; minimum 4s ($0.40), maximum 300s ($30.00).

Fal.aiNiet beschikbaar

OmniHuman 2.0 is not listed on Fal.ai.

ReplicateNiet beschikbaar

OmniHuman 2.0 is not available on Replicate.

** Prijzen van concurrenten worden geschat op basis van vergelijkbare modelarchitecturen en gebruiksniveaus.

⚙️

Technical Details

Configuration schema

snelstring

Optional prompt (max 300 characters) for scene, character motion and camera movement.

Default ValueA museum guide explains how starlight lets us look into the universe's past in a friendly, engaging manner.
Afbeeldings-URLstring

URL of the portrait image (jpeg, png, webp, bmp, tiff, gif, heic; under 30 MB; aspect ratio 0.4-2.5).

Default Valuehttps://cdn.muapi.ai/assets/omnihuman-1-5.jpg
Audio-URLstring

URL of the audio that drives motion and lip sync (wav or mp3, 4-300 seconds, under 15 MB).

Default Valuehttps://cdn.muapi.ai/assets/omnihuman-1-5.mp3
Watermarkboolean

Add a watermark to the output video.

Default Valuefalse
📖

Implementation Guide

Documentatie voor ontwikkelaars

How to Use OmniHuman 2.0

  1. Prepare the portrait: Provide a clear image URL (under 30 MB, aspect ratio between 0.4 and 2.5, 300 to 6000 px per side).

  2. Provide the audio: Use a wav or mp3 URL between 4 and 300 seconds, under 15 MB.

  3. Optional prompt: Add up to 300 characters describing the scene, motion and camera.

  4. Submit and collect: Send the request and poll for the result.

curl -X POST https://api.muapi.ai/api/v1/omnihuman-2-0 \
  -H "x-api-key: $MUAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{"image_url": "https://example.com/portrait.png", "audio_url": "https://example.com/speech.mp3", "prompt": "A museum guide explains how starlight lets us look into the past in a friendly, engaging manner."}'

Then poll GET /api/v1/predictions/{request_id}/result until the status is completed, or pass a webhook query parameter.

❓

Common Questions

Frequently asked

What inputs does OmniHuman 2.0 need?

One portrait image (image_url) and one audio file (audio_url). A text prompt (up to 300 characters) is optional.

How long can the audio be?

Between 4 and 300 seconds. Audio must be wav or mp3 and under 15 MB.

How is the cost calculated?

$0.10 per second of audio, rounded up to the next whole second, with a 4-second minimum (a 4s clip costs $0.40) and a 300-second maximum ($30.00).

Does the output include a watermark?

No by default. Set enable_watermark to true to add one.

Which languages can the prompt use?

Chinese, English, Japanese, Korean, Spanish (Mexican) and Indonesian.

How is it different from OmniHuman 1.5?

OmniHuman 2.0 takes longer audio (up to 300 seconds), supports prompt control of scene, motion and camera, and is billed per second of audio.