HumanAIGC-Engineering/OpenAvatarChat

Health v2 · maintenance only

Adoption not part of the score

3723 stars · 620 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

OpenAvatarChat is a modular interactive digital human (talking avatar) chat application that combines ASR, LLM, TTS, and avatar rendering components into a low-latency real-time conversation system. It supports multiple avatar technologies (LiteAvatar, LAM, MuseTalk, FlashHead) with swappable components and ~2.2 second average response time.

Use cases

• build a real-time talking avatar chatbot
• run a digital human assistant that listens and speaks
• swap ASR, LLM, TTS, and avatar components in a voice chat pipeline
• create an interactive virtual presenter with lip-synced video
• demo a multimodal conversational AI with voice and video input
• self-host a low-latency voice-to-avatar conversation service

When to choose

• you need a full-stack digital human chat demo or product with swappable components
• you want real-time voice conversation with a lip-synced avatar
• you want to experiment with different avatar rendering technologies in one framework

When to avoid

• you only need a text-only chatbot without avatars or speech
• you lack a GPU or cannot host models locally
• you need a production-grade managed service rather than a self-hosted research framework

Facets

application · maturity active

chatbot speech-recognition tts llm-inference agent-framework web-framework artificial-intelligence large-language-models chatbots speech-processing computer-vision python self-hosted cross-platform digital-human talking-avatar real-time-conversation asr vad multimodal webrtc modular-pipeline real-time web-server gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="HumanAIGC-Engineering/OpenAvatarChat")