Header Logo

Docs

Start

Header Logo

Start

Header Logo

Docs

Start

All Models

ByteDance logo

OmniHuman Avatar

OmniHuman Avatar creates talking-avatar videos from a portrait and audio, combining synchronized mouth movement with facial expressions and head motion. It supports speech and singing.

Price per input second. Round the total charge up to the next $0.01.

OfferingPrice (USD)
native$0.12
  • Round billed duration up to a whole second before calculating the charge.
  • native: minimum billed duration 1s.

OmniHuman Avatar

OmniHuman Avatar animates a portrait using an audio recording to create a talking-avatar video. The image supplies the subject's appearance, and the recording supplies the vocal performance. Mouth movements follow the audio, accompanied by facial expressions and head motion.

The model supports spoken recordings and singing, while retaining the pictured subject's likeness. This gives creators a way to build a visible performance around an existing voice recording, using a portrait as the character's visual identity.

What it's good for

Training teams can turn recorded explanations into presenter-led segments. A pictured instructor can deliver course narration, while a spokesperson portrait can accompany a product introduction or a personalized message.

For creative work, illustrated characters can perform dialogue or a vocal passage in social videos. Pairing the same portrait with different recordings also supports related presenter segments, with the source image providing a consistent visual reference across the material.

Model ID

bytedance/omnihuman

Output Type

Video

Pricing

$0.12 / input second

Provider

ByteDance

Hosting

Passthru

OmniHuman Avatar

OmniHuman Avatar animates a portrait using an audio recording to create a talking-avatar video. The image supplies the subject's appearance, and the recording supplies the vocal performance. Mouth movements follow the audio, accompanied by facial expressions and head motion.

The model supports spoken recordings and singing, while retaining the pictured subject's likeness. This gives creators a way to build a visible performance around an existing voice recording, using a portrait as the character's visual identity.

What it's good for

Training teams can turn recorded explanations into presenter-led segments. A pictured instructor can deliver course narration, while a spokesperson portrait can accompany a product introduction or a personalized message.

For creative work, illustrated characters can perform dialogue or a vocal passage in social videos. Pairing the same portrait with different recordings also supports related presenter segments, with the source image providing a consistent visual reference across the material.