OmniHuman 1.5
- Clip lengths
- 1–30s
- Resolutions
- 720p
from 140 credits
AI Talking Avatar
Turn a single portrait and your WAV recording into an audio-driven avatar video. Review the model settings, credit quote and generated performance.
This selection requires purchased credits. Welcome credits apply to the Standard Seedance 1.0 Pro Fast preset. Review the current cost before generating.
Bring the portrait and the spoken audio you want to use. This workflow animates one photographed subject from an uploaded WAV recording; it does not create narration from typed text. A clear face and a clean recording give you a more useful starting point for reviewing lip movement and expression.
Concept demonstration to illustrate a creative direction. This is not a verified output from this tool.
Prepare your source files
Single-subject portrait + WAV recording
Review before publishing
Review lip movement, expression and audio timing
From source files to video
Choose a portrait
Upload a clear photo of one subject with the face visible and little obstruction.
Add your recording
Upload your WAV recording using the controls on this page. Choose the model and check the permitted duration and credit quote.
Create and review
Confirm generation on this page, then listen to the audio and inspect lip movement, facial expression and identity.
Explore available models
What you can create
Short introductions
Animate a portrait for a prepared welcome or personal introduction.
Lesson segments
Present a short explanation using audio you have already recorded.
Character dialogue
Explore a fictional character speaking a prepared line with authorized source material.
Questions before you start
Can I type a script to create speech?
This workflow requires an uploaded WAV recording. It does not include text-to-speech or automatic voice creation.
What should I upload?
A single-subject portrait and a supported uncompressed WAV recording. Use the format and duration limits shown by the selected model.
Can I use a group photo?
Choose one clearly visible subject. Group portraits and obstructed faces can make the intended speaker ambiguous.
Is lip sync always exact?
No. Review lip movement, facial detail and audio timing before publishing, especially around fast speech or unusual expressions.