Back to Models Guide
Bytedancevideo

Complete Guide to Using OmniHuman 1.5

Try This ModelTutorial

Overview

Turn a portrait photo and an audio clip into a lifelike talking video with natural motion and accurate lip sync

Input Parameters

image_url
filerequired

Portrait image to animate (JPEG/PNG/WebP, ≤10MB).

audio_url
filerequired

Audio to drive the performance (MP3/WAV/AAC/OGG/MP4, ≤10MB, under 60 seconds — 15s or less recommended).

prompt
textarea

Optional guidance for the performance (max 1000 characters).

Resolution
select

Output resolution.

Options
720p1080p
Default: 1080
Fast Mode
toggle

Speeds up generation at some quality cost.

Default: false
seed
number

Random seed. Positive integers yield consistent results; -1 = random.

Quick Info
ProviderBytedance
Categoryvideo
Available on
lip-sync
Sections
OverviewInput Parameters
Visit Bytedance website