OmniHuman·by ByteDanceTalking avatar

OmniHuman 1.5

ByteDance OmniHuman 1.5: turns a portrait image and an audio track into a realistic, lip-synced talking video with facial expression and body motion. Billed per second of audio (≤60s).

Open in workspacebytedance/avatar-omni-human-v1.5

Parameters

ParameterTypeDefaultRange or options
seed
seed

Random seed; -1 for random.

number-1
Resolution
output_resolution

Output video resolution (720 or 1080). Does not affect price.

select1080720, 1080

Inputs

Files this model takes in addition to the prompt.

InputAcceptsMax files
Image
image_url
image1Required
Audio File
audio_url
audio1Required