ARIA: Adaptive Responsive Intelligence Assistant

🚧 WORK IN PROGRESS - Expect breaking changes, incomplete features, and active development

A privacy-first, multi-user household voice assistant with voice authentication and hands-free navigation

⚠️ Beta Software: This is experimental software under active development. Features may be incomplete, APIs may change, and bugs are expected. Use at your own risk.

ARIA is a self-hosted voice assistant designed for small households (2-5 members) to coordinate calendars, manage tasks, share shopping lists, track budgets, and stay organized—all through natural conversation.

Key Features:

• 🎙️ Voice Authentication: Particle-based visual auth system with voice guidance
• 🗣️ Hands-Free Navigation: Wake-word activation, voice command confirmation
• 🔒 Privacy-First: Self-hosted, zero telemetry, all data stays local
• 🎨 Audio-Reactive UI: Real-time soundwave visualization synced to voice

Features

• Natural Voice Interaction: Wake-word activation, voice command confirmation, neural voice synthesis
• Audio-Reactive UI: Real-time soundwave visualization synced to voice
• Particle-Based Auth: Beautiful animated login/register with voice guidance
• Multi-User Support: Separate personal data, shared household resources
• Self-Hosted & Private: Zero telemetry, all data stays local, you own your data
• Multi-LLM Support: Claude, Gemini, OpenAI, or local models (Ollama)
• Flexible TTS: Azure/Google neural voices or browser Web Speech API

Screenshots

Login	Voice Selection	Dashboard

Tech Stack

• Frontend: React 19 + TypeScript + Vite + TailwindCSS + Three.js
• Backend: Node.js + Express + Prisma (MongoDB)
• Agent: Python + FastAPI + LangChain
• Deployment: Docker Compose (single command setup)

Quick Start

# Clone repository
git clone https://github.com/yourusername/ARIA.git
cd ARIA

# Setup environment
cp .env.example .env
# Edit .env with your API keys (see Configuration section below)

# Launch with Docker
docker-compose up -d

# Access ARIA
# Frontend: http://localhost:3004
# Backend API: http://localhost:5001
# Python Agent: http://localhost:8002

Configuration

Required Environment Variables

Backend (.env in root):

# MongoDB
MONGODB_URI=mongodb://localhost:27019/aria?replicaSet=rs0

# JWT Secrets
JWT_SECRET=your-super-secret-jwt-key-change-this-in-production
JWT_REFRESH_SECRET=your-super-secret-refresh-key-change-this-too

# Redis
REDIS_URL=redis://localhost:6381

Frontend (.env in frontend/):

# API Endpoint
VITE_API_URL=http://localhost:5001

# Azure TTS (Optional - for natural voice synthesis)
VITE_AZURE_SPEECH_KEY=your-azure-speech-key
VITE_AZURE_SPEECH_REGION=eastus

# Voice Settings (Fallback to Web Speech API if Azure not configured)
VITE_VOICE_ENABLED=true

Azure TTS Setup (Optional but Recommended)

Azure Cognitive Services Text-to-Speech provides natural, human-like voice for ARIA instead of robotic Web Speech API.

Free Tier: 5 million characters/month (~16,000 interactions)

Setup Steps:

1. Create Azure Account: https://azure.microsoft.com/free/
2. Create Speech Service:
   • Go to Azure Portal → Create Resource → Search "Speech"
   • Create new Speech Services resource
   • Choose Free (F0) pricing tier
   • Select region (e.g., eastus)
3. Get Credentials:
   • Go to your Speech resource → Keys and Endpoint
   • Copy KEY 1 and LOCATION/REGION
4. Add to .env:

   VITE_AZURE_SPEECH_KEY=your-key-here
   VITE_AZURE_SPEECH_REGION=eastus

Without Azure: ARIA automatically falls back to browser's Web Speech API (works but sounds robotic).

Development

ARIA is actively developed. See Project Plan for detailed roadmap.

License

PolyForm Noncommercial License 1.0.0

Copyright (c) 2025 Phillip-Juan van der Berg

• ✅ Free for personal use, learning, and research
• ✅ Free for nonprofits, educational institutions, and open source projects
• ✅ Modify and distribute with attribution
• ❌ Commercial use forbidden without explicit permission

What is Commercial Use?

• Using ARIA in a business or for-profit operation
• Selling access to ARIA as a service
• Integrating ARIA into commercial products
• Using ARIA to generate revenue

Commercial Licensing: Contact GitHub for commercial licensing options.

See LICENSE file for full legal text.

Testing

# Frontend (React + Vitest)
cd frontend && npm test

# Backend (Node + Vitest)  
cd backend && npm test

# Agent (Python + Pytest)
cd agent && pip install -r requirements.txt && pytest

Contributing

See Core Rules for development standards. Please open an issue to discuss major changes before submitting a PR.