Project AURA (Advanced Universal Responsive Avatar)

AURA is an open-source interactive AI Avatar project developed by the Advanced Software Engineering (ASE) Laboratory. Inspired by AI VTubers, AURA serves as a virtual assistant capable of answering questions, managing lab information, and providing conversational support via a web portal.

🚀 Features

Conversational Core

• Multi-modal Interaction: Supports both text-based chat and low-latency voice interaction.
• Local AI Edge: Capable of running high-performance TTS (Qwen3) entirely locally on consumer GPUs.
• Polyglot Intelligence: Automatically detects and responds in English, Indonesian, and Japanese.
• RAG Memory: Integrated with Supabase (pgvector) to remember past conversations and lab-specific knowledge.
• LiveKit Integration: Uses LiveKit Agents for high-performance audio streaming and VAD (Voice Activity Detection).

Dashboard

• Modern React-based dashboard for chat, personality management, and call modes.

🛠️ Tech Stack

• Orchestration: Python, FastAPI, LangGraph.
• Voice Stack: LiveKit, Deepgram (STT), Qwen3-TTS (Local) / Cartesia (Cloud).
• LLM: OpenRouter (Unified API for DeepSeek, GPT-4, etc.).
• Database: Supabase (PostgreSQL + pgvector).
• Frontend: React, Vite, TailwindCSS.
• Deployment: Windows Launcher (Native) & Docker / Docker Compose.

⚡ Quick Start

Prerequisites

• Docker & Docker Compose

Installation

1. Clone the repository:

   git clone https://github.com/ASE-Lab/project-aura.git
   cd project-aura

2. Configure Environment: Copy the example environment file and fill in your API keys.

   cp .env.example .env

   Note: Set TTS_TYPE=qwen for local or cartesia for cloud.

3. Run the Launcher (Windows Recommended):

   ./start_aura.bat

   This will handle environment checks and start all services (Dashboard, AI Service, Voice Agent, Token Server).

Alternative: Docker Setup

docker compose up --build -d

🌐 Accessing Services

• User Dashboard: http://localhost:5173
• AI Service Docs: http://localhost:8000/docs
• Voice Agent: Integrates automatically with the Dashbord.

🤝 Contribution

We follow a standard open-source contribution workflow!

1. Fork the repository.
2. Create a Branch for your feature (git checkout -b feat/your-feature).
3. Commit your changes.
4. Push to the branch.
5. Open a Pull Request.

📜 License

Distributed under the MIT License. See LICENSE for more information.

---

Built with ❤️ by the ASE Lab Members