ClusterTalk is a production-grade, distributed real-time messaging platform engineered with a microservices architecture. It demonstrates mastery of distributed systems, cross-platform engineering, and applied AI integration.

Designed for high-throughput and low-latency, it seamlessly orchestrates a Go websocket engine, Python AI moderation service, Rust analytics consumer, and React/Mobile clients.

graph TD
    subgraph Clients
        Web[React Web Client]
        Android[Android App (Kotlin)]
        iOS[iOS App (Swift)]
    end

    subgraph "Core Cluster"
        LB[Load Balancer / Gateway]
        Hub[Go WebSocket Hub]
    end

    subgraph "Microservices"
        AI[Python AI Service]
        Analytics[Rust Analytics Engine]
    end

    Web <-->|WebSocket| Hub
    Android <-->|WebSocket| Hub
    iOS <-->|WebSocket| Hub
    
    Hub -->|HTTP Sync| AI
    Hub -.->|HTTP Async| Analytics

• 🚀 High-Performance Core: Go-based event loop using lightweight goroutines handling thousands of concurrent connections.

• 🛡️ AI-Powered Moderation: Real-time toxicity detection using a Python Microservice. Automatically censors spam and offensive content before broadcast.

• 💾 SQLite Persistence: ACID-compliant chat history storage using a pure Go driver. Replays recent messages upon joining a room.

• 📊 Rust Analytics: Zero-overhead event ingestion pipeline implemented in Rust for analyzing chat throughput.

• 🌐 Multi-Room Clustering: logical partitioning of chat streams (#general, #tech, #random).

• 📱 Cross-Platform Architecture: Unified protocol supporting Web, Android, and iOS clients.

• 🎨 Cyberpunk UI: Modern, glassmorphism-based interface with responsive design.

• Go 1.19+
• Node.js 16+
• Python 3.9+
• Rust (Cargo)

The heart of the system.

cd backend-go
go mod tidy
go run cmd/server/main.go
# 🟢 Server listening on :8080

The brain of the system.

cd ai-python
# Create virtual env
python -m venv venv
./venv/Scripts/activate
pip install -r requirements.txt
# Start Service
uvicorn main:app --reload --port 8000
# 🟢 AI Service online at :8000

The face of the system.

cd client-web
npm install
npm run dev
# 🟢 Client ready at http://localhost:5173

The memory of the system.

cd analytics-rust
cargo run
# 🟢 Analytics ingestion at :9000

1. Open Client: Navigate to http://localhost:5173.
2. Join Room: You are automatically placed in #general.
3. Test Chat: Send a message. It appears instantly.
4. Test AI: Send the message "This is spam". It will be intercepted and replaced with [CENSORED BY AI].
5. Multi-Room: Open a second tab, switch to #tech. Messages are isolated between rooms.

ClusterTalk/
├── backend-go/         # Go WebSocket Server
├── ai-python/          # Python AI Microservice
├── analytics-rust/     # Rust Data Consumer
├── client-web/         # React Frontend
├── client-android/     # Native Android Project
├── client-ios/         # Native iOS Project
└── infra/              # Docker & Kubernetes Configs

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