human  by LKZMuZiLi

Unity-based real-time 3D avatar

created 1 year ago
370 stars

Top 77.5% on sourcepulse

GitHubView on GitHub
Project Summary

This project provides an open-source, real-time 3D digital human powered by Unity, targeting developers and researchers interested in creating interactive AI-driven characters. It integrates speech recognition, LLM-based conversational AI, and text-to-speech with lip-syncing for a lifelike experience.

How It Works

The system processes user microphone input through speech recognition, feeds the text to a chosen LLM API for response generation, and then uses text-to-speech (TTS) for audio output. Lip synchronization is achieved using the uLipSync package, ensuring mouth movements match the synthesized speech. The architecture supports various LLM APIs and is built on Unity's Universal Render Pipeline (URP) for cross-platform compatibility.

Quick Start & Requirements

  • Install: Requires Unity 2022.3.31 or later.
  • Prerequisites: LLM API key and configuration (URL, model name, optional role settings). Integration with the Fay digital human framework is supported.
  • Resources: Downloadable runtime package available via Baidu Netdisk.
  • Documentation: Reference video available at https://www.bilibili.com/video/BV1YJ4m1T7Rj?t=2.7.

Highlighted Details

  • Utilizes URP for rendering and supports all platform releases.
  • Integrates with various LLM APIs (e.g., Doubao, Zhipu, Kimi).
  • Features uLipSync for accurate lip-syncing.
  • Designed for extensibility and ease of use.

Maintenance & Community

The project is maintained by LKZMuZiLi. Community engagement is encouraged via WeChat for group invitations.

Licensing & Compatibility

The specific license is not explicitly stated in the README, but the project is presented as an open-source version derived from a primary project. Compatibility for commercial use or closed-source linking would require clarification on licensing terms.

Limitations & Caveats

Users must configure LLM API details themselves. Older Unity versions may require manual handling of URP package errors. The project relies on external services for LLM and TTS, and the availability and performance of these services are external dependencies.

Health Check
Last commit

4 months ago

Responsiveness

1 day

Pull Requests (30d)
0
Issues (30d)
1
Star History
53 stars in the last 90 days

Explore Similar Projects

Starred by Thomas Wolf Thomas Wolf(Cofounder of Hugging Face), Chip Huyen Chip Huyen(Author of AI Engineering, Designing Machine Learning Systems), and
2 more.

ultravox by fixie-ai

0.4%
4k
Multimodal LLM for real-time voice interactions
created 1 year ago
updated 4 days ago
Feedback? Help us improve.