xszyou/Fay

What is it?

Key features

• 01Connects digital humans (2.5D, 3D, mobile, PC, web) with LLMs and voice models.
• 02Supports OpenAI-compatible LLM APIs and DeepSeek thinking models.
• 03Lets you swap ASR and TTS models independently.
• 04Provides text, voice, driving, control, broadcast, and intent APIs.
• 05Supports wake-word, interruption, and multi-user concurrent sessions.
• 06Offers custom knowledge bases, Q&A pairs, persona settings, and MCP tool management.

Use cases

• Building a virtual teacher, virtual host, or automated news broadcaster.
• Adding a voice-driven digital human assistant to a website or app.
• Connecting an LLM to an existing business system through a unified API.
• Running an offline voice and avatar assistant on local hardware.

Quick start

1. 1

   Install Python 3.12 and, on Ubuntu, build-essential and portaudio19-dev

2. 2

   Install dependencies from requirements.txt

   pip install -r requirements.txt
3. 3

   Start the project with the public config center (slow; replace with your own key)

   python main.py start -config_center d19f7b0a-2b8a-4503-8c0d-1a587b90eb69
4. 4

   Open the management page in a browser at http://127.0.0.1:5000

Things to know

• !Requires Python 3.12 specifically.
• !The default public config center is very slow, so you should configure your own API key.
• !Documentation and community resources are mainly hosted on Feishu and in Chinese.

Topics

FAQ

What is Fay?

An agent framework that connects digital humans (2.5D, 3D, mobile, PC, web) and LLMs (OpenAI-compatible, DeepSeek) to business systems. Fay is a framework for putting digital humans (avatars) into real products. It adapts to different digital human rendering technologies on the front end and connects to large language models, including OpenAI-compatible APIs and DeepSeek, on the back end. Speech recognition (ASR) and text-to-speech (TTS) models can be swapped out, so developers can assemble the voice and avatar pipeline they need.

Is Fay free to use? What license does it have?

Fay is open source under the GPL-3.0 license. Check the license terms before using it in commercial products.

Is Fay still maintained?

Actively maintained. The last commit to Fay was on Sep 21, 2026.

What can I use Fay for?

Building a virtual teacher, virtual host, or automated news broadcaster. Adding a voice-driven digital human assistant to a website or app. Connecting an LLM to an existing business system through a unified API. Running an offline voice and avatar assistant on local hardware.

How do I install Fay?

Install Python 3.12 and, on Ubuntu, build-essential and portaudio19-dev Install dependencies from requirements.txt: pip install -r requirements.txt Start the project with the public config center (slow; replace with your own key): python main.py start -config_center d19f7b0a-2b8a-4503-8c0d-1a587b90eb69 Open the management page in a browser at http://127.0.0.1:5000

Where can I get Fay?

The source code and installation instructions are on GitHub: https://github.com/xszyou/Fay