Breeze

AI-powered messaging and calling platform with an agent that can act on your behalf.

Breeze blends real-time chat, voice and video calls, and an agentic LLM assistant that can summarize conversations, enhance tone, schedule reminders, and send messages for you. It is built as a full-stack product with a NestJS backend, a modern React frontend, and security-first session management.

Live Demo

Live URL: https://breeze-latest.onrender.com/

Key Features

Agentic AI Assistance

• Summarize conversations on demand
• Enhance messages with tone control and rewrites
• Schedule reminders with confirmation and delivery
• AI message writer jobs to draft and send messages on your behalf

Messaging and Calling

• Real-time 1:1 and group chat
• Voice and video calls with live signaling
• Presence, typing indicators, and notifications
• Emoji support and rich message attachments

Media and Storage

• S3-backed uploads for voice notes and media
• Presigned URL workflows for secure uploads
• Avatar and asset storage

Security and Sessions

• JWT access and refresh token rotation
• Redis-backed access token blacklist for instant logout
• Anomaly detection with step-up authentication
• Session family management and revocation endpoints

Infrastructure

• WebSocket transport via Socket.IO
• Redis integration for cache and security features
• Postgres persistence with migrations
• Scheduler-driven background processing

Tech Stack

Frontend

Technology	Purpose
React 19 + TypeScript	UI framework
TanStack Start + Router	App routing and server tooling
Vite	Build tool and dev server
Tailwind CSS + Radix UI	Styling and components
Socket.IO Client	Real-time messaging and calls

Backend

Technology	Purpose
NestJS + TypeScript	API and realtime server
TypeORM	Database ORM and migrations
PostgreSQL	Primary database
Redis	Token blacklist and cache
Socket.IO	WebSocket transport
OpenAI-compatible SDK	LLM integration
AWS S3	Media storage and presigned uploads

Project Structure

Breeze/
├── backend/                 # NestJS API and realtime services
├── frontend/                # React UI (TanStack Start)
├── nginx/                   # Reverse proxy config
├── docker-compose.dev.yml   # Local dev services
├── Dockerfile
└── start.sh

Getting Started

Prerequisites

• Node.js 20+
• PostgreSQL
• Redis (Docker recommended)
• An OpenAI-compatible API key (or GitHub Models access)
• AWS S3 credentials for media uploads

Local Development

1) Start Redis

docker compose -f docker-compose.dev.yml up -d redis

2) Backend

cd backend
npm install
cp .env.example .env
npm run start:dev

3) Frontend

cd frontend
npm install
cp .env.example .env
npm run dev

Default ports:

• Backend: http://localhost:3000
• Frontend: http://localhost:5173

Environment Variables

Key backend variables live in backend/.env.example. Highlights:

• DATABASE_URL
• JWT_ACCESS_SECRET, JWT_REFRESH_SECRET
• REDIS_HOST, REDIS_PORT
• S3_BUCKET, S3_REGION, S3_ACCESS_KEY_ID, S3_SECRET_ACCESS_KEY
• AI_MODEL, AI_BASE_URL, GITHUB_MODEL_KEY

Frontend variables live in frontend/.env.example:

• VITE_API_URL
• VITE_VAPID_PUBLIC_KEY

Documentation

• Anomaly detection system
• Access token blacklist
• Refresh event logging
• Auth flow diagrams

License

This project is open source and available under the MIT License.