Characters that run anywhere
bitHuman builds on-device, real-time avatar models you can deploy anywhere — from a Raspberry Pi to a GPU, offline or in the cloud, behind a drop-in SDK.
- Raspberry Pi → GPU
- runs on the hardware you own
- Real-time
- on-device, offline or in the cloud
- Drop-in SDK
- Python and Swift
- Linux · macOS · iOS · edge
- deploy anywhere
Why we build this
Every interface deserves a face. We make avatars that run where your users are — on their device, at the edge, offline — not just in someone else's cloud.
What we build
Two avatar models for different hardware budgets, and the SDK, API, and dashboard that put them wherever you deploy.
Essence
Lightweight, on-deviceAn avatar that runs in real time on everyday CPUs — Raspberry Pi to cloud, no GPU required.
Expression
Expressive, full fidelityHigh-fidelity neural rendering in real time on NVIDIA GPUs and Apple Silicon.
Python SDK
Inference, sessions, and LiveKit integration on PyPI.
Swift SDK
On-device voice pipeline for iOS and macOS via SwiftPM.
Cloud API
Create agents and run sessions — no infrastructure to manage.
Web Dashboard
Build agents, embed chat, and monitor usage — no code.
Open by default
Open SDKs and examples on PyPI and SwiftPM — from install to a running avatar in minutes.