Bithuman

1.10.8 verified Fri May 01 auth: no python

Real-time avatar engine (v1.10.8) generating lip-synced video and streaming live avatars to browsers at 100+ FPS on CPU. Requires Python >=3.9, supports ARM/x86/macOS. Active development with regular releases.

pip install bithuman

Common errors

error ModuleNotFoundError: No module named 'bithuman' ↓

cause Package not installed or virtual environment not activated.

fix

Run: pip install bithuman

error AttributeError: module 'bithuman' has no attribute 'Client' ↓

cause Using old class name 'Client' which was renamed in v1.5.

fix

Use: from bithuman import BithumanClient

Warnings

breaking v1.5: 'Client' class renamed to 'BithumanClient'. Breakage if using old name. ↓

fix Update import: from bithuman import BithumanClient

deprecated v1.6: 'AvatarConfig' was moved from bithuman.config to bithuman. Old import deprecated. ↓

fix Use: from bithuman import AvatarConfig

gotcha The 'generate_video' method requires a 'config' argument; passing just text fails silently with AttributeError. ↓

fix Always create an AvatarConfig object and pass it as config=config

Imports

BithumanClient

wrong

from bithuman import Client

correct

from bithuman import BithumanClient

Old SDK used 'Client'; renamed in v1.5

AvatarConfig

wrong

from bithuman.config import AvatarConfig

correct

from bithuman import AvatarConfig

Config moved to top-level in v1.6

Quickstart

Initialize client and generate a lip-synced video from text.

from bithuman import BithumanClient, AvatarConfig
import os

client = BithumanClient(api_key=os.environ.get('BITHUMAN_API_KEY', ''))
config = AvatarConfig(avatar_id='default', language='en')
result = client.generate_video(text='Hello, world!', config=config)
print(result.video_url)