Create AI avatar videos from a single image
Turn one image into a realistic AI avatar video. Combine the image and audio to create natural facial expressions, head movement, and lifelike animation.
OmniHuman 1.5
OmniHuman 1.5 is ByteDance's AI avatar and digital human model. It turns a single image and an audio clip into a realistic talking video with natural lip sync, expressive faces, and lifelike movement. One image and audio are enough to create avatar, lip-synced, and digital human videos.
Create now
OMNIHUMAN 1.5 STUDIO
Bring a portrait to life with audio
Upload one image and an audio clip, choose the output, review the credit quote, then create your video.
Portrait image
Driving audio
For multiple subjects, upload a mask for the one you want to animate.
0/300
-1 is random; reuse a non-negative number for greater consistency.
Your creations
Follow your OmniHuman 1.5 tasks and revisit finished videos.
Loading creations…
OmniHuman 1.5 core capabilities
Create realistic AI avatar, lip-synced, and digital human videos from one image and an audio clip.
Create precise lip-synced avatar videos with OmniHuman 1.5
Match speech to mouth movement so the subject speaks naturally with realistic lips and expressive facial details.
Create emotion-aware OmniHuman 1.5 digital human videos
Use the meaning and tone of the audio to guide matching emotions, gestures, and body language for a more engaging video.
Create full-body AI avatar animation with OmniHuman 1.5
Go beyond a traditional talking-photo tool with natural upper-body and full-body movement.
Use OmniHuman 1.5 in AI video applications
Use the model for digital human generators, virtual idols, avatar platforms, AI video tools, and creative products.
Image, audio, and video output
Prepare a clear image and an audio clip before creating a digital human video.
| Item | Media or format | Guidance |
|---|---|---|
| Image | JPG/JPEG, PNG, or WebP; up to 10 MB | A well-lit, front-facing image with a clear face is recommended. Pets and anime subjects are also supported. |
| Audio | MP3, WAV, AAC, OGG, or MP4 audio; up to 10 MB | Audio must be under 60 seconds. For more consistent results, 15 seconds or less is recommended. |
| Video | MP4 at 720P or 1080P | The generated subject's movement follows the supplied audio. |
From one image to an avatar video
Prepare an image, add audio, and review the synchronized result.
- 01
Upload a clear image. A well-lit portrait with visible facial details works best.
- 02
Add spoken, sung, or recorded audio to guide the mouth, expressions, and body movement.
- 03
Generate and review the synchronized MP4 digital human video.
Where OmniHuman 1.5 fits
Use an image and audio for virtual creators, talking photos, music, and brand content.
Virtual creators and influencers
Use a single image and audio file to create realistic virtual influencer content and engaging short videos for creators, agencies, and brands.
Lip-synced avatar videos
Bring natural speech synchronization to talking-photo experiences, avatar generators, virtual presenters, and personalized videos.
AI music videos
Turn songs, recordings, and audio tracks into expressive music videos with precise lip sync and emotion-rich performances.
Digital spokesperson videos
Create brand spokespeople, product explainers, and promotional videos while keeping the character consistent across campaigns.
OmniHuman 1.5 FAQ
What is OmniHuman 1.5?+
OmniHuman 1.5 is ByteDance's AI avatar and digital human model. It turns one image and audio into a video with synchronized lips, facial expressions, and natural movement.
How does it create a digital human video?+
Provide one image and an audio clip. The model analyzes the voice, emotion, and subject to generate a video synchronized with the audio.
Can it create lip-synced videos?+
Yes. OmniHuman 1.5 matches mouth movement to speech while creating natural facial expressions and movement.
Which image formats are supported?+
JPG/JPEG, PNG, and WebP images up to 10 MB are supported. A clear, front-facing image is recommended.
How long should the audio be?+
Audio of 15 seconds or less is recommended. It must be shorter than 60 seconds; longer clips may reduce visual quality.
Can it animate anime or cartoon characters?+
Yes. The model supports people, pets, anime, cartoon characters, and stylized subjects.
Can it show upper-body or full-body movement?+
Yes. In addition to facial and head movement, OmniHuman 1.5 supports natural upper-body and full-body animation.
What format is the video?+
The generated video is an MP4 file suitable for social, web, and content-creation workflows.
See how digital human videos are made
OmniHuman 1.5 is ByteDance's AI avatar and digital human model. It turns a single image and an audio clip into a realistic talking video with natural lip sync, expressive faces, and lifelike movement. One image and audio are enough to create avatar, lip-synced, and digital human videos.
Create now