Architecture

Pump Guard is built as a lightweight, modular system that analyzes tokens in real-time and delivers results instantly to users.

​

System Components

​

Frontend (React + Vite)

• Token search and address validation
• Live risk reports with visual metrics
• Real-time updates via WebSockets
• Recent scans tracking with local storage

​

Backend (Node.js + Express)

• REST API for token analysis
• Solana RPC integration for blockchain data
• DexScreener API for market data
• Redis caching for performance
• Socket.io for live updates

​

Infrastructure

• Solana RPC: Fetches holder balances and wallet history
• Redis: Caches wallet data (24h) and risk scores (15s)
• WebSockets: Broadcasts scan events to all connected users

​

How a Scan Works

1. User enters token address on pumpguard.xaxios.com
2. Backend validates address and checks DexScreener for market data
3. System fetches top 50 holders from Solana blockchain
4. Each wallet analyzed: creation date, SOL balance, token amount
5. Risk metrics calculated using our scoring algorithm
6. Results cached and returned to user
7. Scan event broadcast to all connected clients

​

Performance Strategy

Caching:

• Wallet creation data cached 24 hours (rarely changes)
• Risk scores cached 15 seconds (balance between freshness and speed)
• Token metadata cached indefinitely (immutable)

Rate Limiting:

• 2 requests per second per IP
• 10 requests per 15 seconds
• 30 requests per minute

Real-time Updates:

• Socket.io broadcasts new scans globally
• Homepage shows community scan activity live
• No page refresh needed

​

Tech Stack

Backend:

• Express.js + TypeScript
• @solana/web3.js for blockchain
• Redis (ioredis) for caching
• Socket.io for WebSockets
• Axios for external APIs

Frontend:

• React 18 + TypeScript
• Vite for fast builds
• Tailwind CSS for styling
• Zustand for state management
• Socket.io-client for live updates

​

Open Source

The entire codebase is open source on GitHub:

• Full backend and frontend code
• Deployment configurations
• API documentation
• Risk algorithm implementation

Anyone can run their own instance, verify the scoring logic, or contribute improvements.

---

Note: This is a community-focused tool. For production use cases requiring higher throughput or custom metrics, the architecture can be extended with dedicated RPC nodes, gRPC streams, and advanced indexing.