Header Logo

Docs

Start

Header Logo

Start

Header Logo

Docs

Start

All Models

ByteDance logo

OmniHuman 2.0 Avatar

OmniHuman 2.0 Avatar makes talking-avatar videos from portraits and recorded speech, with optional text guidance for the character's movement, expression, and performance style.

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

OfferingPrice (USD)
native$0.10
  • Round billed duration up to a whole second before calculating the charge.
  • native: minimum billed duration 4s.

OmniHuman 2.0 Avatar

OmniHuman 2.0 Avatar generates a talking-avatar video using a portrait and an audio recording. The portrait establishes who appears on screen, while the recording determines the speaking performance and timing. Optional text guidance adds direction for movement, expression, or overall delivery style.

These inputs let creators combine a chosen visual identity with recorded narration and performance direction. The voice remains the basis of the spoken delivery, with text providing an additional way to shape how the avatar performs.

What it's good for

Product teams can turn explanatory narration into a presenter-led demonstration or introduction. Course authors can pair instructional recordings with a pictured instructor for lessons and training segments.

It also suits localized presenter content: the same portrait can accompany recordings in different languages or revised scripts. For social and editorial work, existing portrait assets can become speaking characters, with performance guidance helping establish the intended tone of each segment.

Model ID

bytedance/omnihuman-2.0

Output Type

Video

Pricing

$0.10 / input second

Provider

ByteDance

Hosting

Passthru

OmniHuman 2.0 Avatar

OmniHuman 2.0 Avatar generates a talking-avatar video using a portrait and an audio recording. The portrait establishes who appears on screen, while the recording determines the speaking performance and timing. Optional text guidance adds direction for movement, expression, or overall delivery style.

These inputs let creators combine a chosen visual identity with recorded narration and performance direction. The voice remains the basis of the spoken delivery, with text providing an additional way to shape how the avatar performs.

What it's good for

Product teams can turn explanatory narration into a presenter-led demonstration or introduction. Course authors can pair instructional recordings with a pictured instructor for lessons and training segments.

It also suits localized presenter content: the same portrait can accompany recordings in different languages or revised scripts. For social and editorial work, existing portrait assets can become speaking characters, with performance guidance helping establish the intended tone of each segment.