Download
Fay is an open source framework designed to build and deploy interactive digital humans powered by large language models. It acts as a middleware layer that connects digital character technologies with conversational AI systems and business applications. Fay supports various types of digital humans, including 2.5D and 3D avatars, and can be integrated with applications running on mobile devices, PCs, web platforms, and embedded systems. Its architecture allows developers to combine different AI components such as speech recognition, text-to-speech, and large language models to create conversational digital agents. Fay provides multiple interfaces for text, voice, and digital human control, enabling developers to build interactive assistants, virtual presenters, or automated service agents. It also supports custom knowledge bases and configurable behaviors so developers can tailor the personality and responses of the digital human.
Features
- Modular framework that integrates digital human models with large language models
- Supports text chat, voice interaction, and automated digital human broadcasting
- Compatible with interchangeable AI components such as ASR and TTS systems
- Allows custom knowledge bases, Q&A pairs, and configurable personas
- Multi-user concurrent interaction with streaming response support
- Flexible deployment across apps, embedded systems, and web environments
Project Samples
