OpenTalking: A Real-Time AI Digital Human Conversation Framework

Our lab has officially open-sourced OpenTalking, a real-time AI digital human conversation framework for interactive avatar applications.
OpenTalking connects LLMs, speech recognition, text-to-speech, WebRTC audio/video streaming, and multiple digital human driving models into one flexible and extensible stack. It supports real-time conversations, video generation, video cloning, and private deployment, making it suitable for both quick demos and production-oriented digital human applications.
We welcome developers, researchers, and builders interested in digital humans, voice interaction, and multimodal AI to try it out, star the project, open issues, and contribute to the community.
GitHub:
https://github.com/datascale-ai/opentalking