Compare models

Essence 2 renders a photoreal person; Expression 2 renders any character.

A bitHuman model turns speech audio into a lip-synced talking video of your avatar, live (how it works). Pick by subject: a real person, Essence 2; anything else, Expression 2. The same agent works on every platform that runs its model; on your own hardware it renders from an avatar file (.imx).

The models

Essence 2NewHotA photoreal person from one portrait, live.Photoreal, up to 1080pMade for always-on kiosks and displaysRuns on iPhone & iPadMacAndroidLinux (no GPU)BrowserYour serversCloud
Expression 2NewHotAny character from one portrait: stylized, animal, robot or human.Cartoons, animals, robots and peopleMotion generated live from the audioOne photo is enoughRuns on iPhone & iPadMacAndroidLinux (no GPU)BrowserYour serversCloudOffline
Essence 1First generation. A pre-built identity with its mouth matched to the audio in real time.Renders on a CPU, no GPU neededLip-sync patched live over pre-rendered motionCustom gesturesRuns on MacLinux (no GPU)BrowserYour serversCloudOffline
Expression 1First generation. Animates a face from a portrait at runtime.No per-identity build stepRenders on bitHuman cloud GPUs at 512×512Can animate a photo with no agentRuns on Cloud

Essence 1 and Expression 1 are the first generation. They stay supported, and nothing changes for agents that use them.

Choosing a model

Pick the model from the subject:

A real person: Essence 2.
Anything else (a cartoon or stylized character, an animal, a robot, a creature, a whole generated scene): Expression 2.
Not sure: create with model: "auto". A real person routes to Essence 2, anything else to Expression 2.
On a phone, a Mac or in a browser: Essence 2 or Expression 2.
Maintaining a first-generation agent: keep it. Essence 1 runs on your own CPU; Expression 1 runs in the bitHuman cloud.

Every model except Expression 2 needs a clear, real human face; otherwise creation is refused with 422 MODEL_SUBJECT_MISMATCH before anything is charged.

Where each model runs

Each place links to the page that sets it up.

Where	Essence 2	Expression 2	Essence 1	Expression 1
iPhone and iPad	Yes Swift package, `Essence2Kit` (iOS 26)	Yes Swift package, `Expression2`	—	—
Mac	Yes Swift package (macOS 26, M3 or newer), CLI, Python	Yes Swift package (macOS 13), CLI, Python	Yes CLI (`run`), Python	—
Android	Yes `essence2-android`	Yes `expression2-android`	—	—
Linux, no GPU	Yes CLI, Python	Yes CLI, Python	Yes CLI (`run`), Python	—
Browser (WebGPU)	Yes `render=local`, for avatars that have a browser build	Yes `render=local`	Yes `render=local`	—
Your servers	Yes CLI, Python, LiveKit plugin	Yes CLI, Python, LiveKit plugin	Yes CLI (`run`), Python	—
bitHuman cloud	Yes web embed, REST API, LiveKit	Yes web embed, REST API, LiveKit	Yes web embed, REST API, LiveKit	Yes web embed, REST API, LiveKit
Fully offline	Not yet Coming later	Yes Linux x86_64 (Business & Enterprise)	Yes Linux x86_64 and ARM64, macOS on Apple silicon (Business & Enterprise)	—

Fully offline is for Business and Enterprise clients, bought in the Console or through sales (Fully offline).

Rendering on your own hardware, every place but the bitHuman cloud, bills at the self-hosted rate (pricing). How fast each model renders on each device: Performance.