A developer built a hybrid P2P file sharing system using FastAPI and WebSockets for a LAN collaboration project. The architecture splits tasks: a central server handles user discovery and messaging, while files transfer directly between peers. This approach avoids server bottlenecks and improves privacy for local network transfers.

The system uses JWT authentication for secure WebSocket connections and an allowlist for file access. Clients run a local HTTP server for sharing, while the central server only routes metadata. This design mirrors LAN party file sharing, offering faster transfers than traditional centralized cloud storage.

The project highlights practical FastAPI learning, though the AI-generated GUI client remains untested. Future improvements could include encryption and NAT traversal for wider use. It demonstrates how separating coordination from data transfer creates efficient, scalable local tools for developers.