ByteDance · DXBuilder
OmniHuman 2.0
Image + audio, brought to life.
Animate a portrait with your own voice or song. One image, one audio file, up to 300 seconds at 720p.
Open studio ↗From input to video
Choose your files and direction, review the price, then generate. Results remain available in your video history.
- Upload image
- Upload audio
- See the whole price first. → Generate video
Know your creative space.
JPEG/PNG/WebP. One image. OmniHuman: below 30 MB, 300–6000 px, ratio 0.4–2.5. PixVerse: below 20 MB.
Duration follows the uploaded audio (4–300s). MP3/WAV, below 15 MB.
Use English, Chinese, Japanese, Korean, Mexican Spanish or Indonesian for OmniHuman direction. The audio supplies the speech.
How are credits calculated?
Credits are calculated for the chosen duration, resolution and audio option and shown before generation. OmniHuman uses the measured audio duration.
| Resolution | Upload audio |
|---|---|
| 720p | 105 |
Before you press generate.
How does audio work?
OmniHuman uses your uploaded audio. PixVerse generates new audio when enabled; it does not accept an audio reference here.
How are credits calculated?
Credits are calculated for the chosen duration, resolution and audio option and shown before generation. OmniHuman uses the measured audio duration.
Does this guarantee a faithful result?
No. Review the face, words, actions, ending and sound. A completed generation is not proof of fidelity.