Essence 2

Essence 2 renders a photoreal person from one portrait, live.

On the deviceYour computer or serverbitHuman cloud

What it is

Essence 2 (essence-2) renders a photoreal person from one portrait, up to 1080×1920 (portrait). From your portrait the platform generates a 10-second video of that person; the model then animates lip-sync and expression over it live, with a sharp mouth and teeth taken from that video. Creating an Essence 2 avatar takes about 2 hours (Create your own avatar).

How it is created, served and kept moving while idle: How it works.

When to choose it

A photorealistic person — start here.
Always-on displays — kiosks, lobby screens and 24/7 assistants.
On your own hardware — every SDK platform runs it.

Essence 2 needs a clear, real human face: a cartoon, animal, robot or creature, or a photo with no face found, is refused with 422 MODEL_SUBJECT_MISMATCH before anything is charged. For a stylized character, or a scene generated from one photo, choose Expression 2. See Choosing a model.

Where it runs

Where	Essence 2	How
iPhone and iPad	Yes	Swift package, `Essence2Kit` (iOS 26)
Mac	Yes	Swift package (macOS 26, M3 or newer), CLI, Python
Android	Yes	`essence2-android`
Linux, no GPU	Yes	CLI, Python
Browser (WebGPU)	Yes	`render=local`, for avatars that have a browser build
Your servers	Yes	CLI, Python, LiveKit plugin
bitHuman cloud	Yes	web embed, REST API, LiveKit
Fully offline	Yes	Linux x86_64 (Business & Enterprise)

A complete app for iPhone and iPad is the iOS Essence 2 example; for Android, the Android Essence 2 example.

The file you download is <CODE>.imx, from GET /v1/agent/{code}/model/download?model=essence-2 or bithuman pull <CODE> --model essence-2. How fast it renders on each device is on performance.

Limits and expectations

Output plays at 25 frames a second everywhere it runs. How fast a platform renders is on performance.
The downloadable file is about 140–160 MB, varying per identity — read Content-Length rather than assuming a size.
The face is fixed at creation. To change it, create a new agent.
The first session on a new agent can take longer to connect while the identity is provisioned; later sessions reuse it.
Before training completes, a launch that requests this model is refused with 409 MODEL_NOT_GENERATED. Once ready, essence-2 appears in the agent’s supported_models.

Next steps

Models — the four models side by side
Agents API — create, poll, download
Embed widget — a live session in minutes
Video API — render an MP4 with model: "essence-2"
Session behavior & troubleshooting