A clear portrait image and an audio track.

Keep it short for best results.

AI Lip Sync VideoHOT

Turn a portrait plus audio into a talking avatar with accurate lip sync.

Lip syncTalking avatarAudio-drivenPromo clips

AI Video Generator

It takes a few runs to find the model and prompt that suit your job; once a version works, you keep reusing it.

Model

Auto-adds quality descriptors to your prompt · off by default

Share your prompt & settings to the Showcase

REFS 0/1

Optional · JPG/PNG · drag images in or click + for image-to-image / image-to-video

Aspect ratio3 common

ResolutionOutput 1024×1024

Duration

Reference audio (lip-sync)

Reference video

Negative prompt (optional)

OmniHuman 1.5 (数字人)·Audio-driven talking avatar

~135 credits this run

Credits are returned if it fails · stay on this page until the result appears

Your generated image will appear here

TRY

Credits are charged per generation by model compute (from ~1 credit).

Models you can use

Switch between them in the model dropdown above.

OmniHuman 1.5 (数字人)

Audio-driven talking avatar

Video720–1080

Wan 2.7 II

Image to video (first frame, optional audio drive), 2–15s, ratio follows source

Video720p–1080p2–15s

Accurate lip sync

Mouth movement matches your audio.

Talking avatars

Great for spokesperson and promo clips.

Audio-driven

Drive a portrait with any voice track.

Commercial safe

No deepfakes of real people without rights.

How it works

1

Upload a portrait image.

2

Upload the voice / audio track.

3

Generate a lip-synced video.

FAQ

It animates a portrait to speak your audio.