How I Built a Lightweight AI Identity Engine on a Cheap 8 GB VPS
I recently launched AnonID, a self-hosted, privacy-focused alternative to commercial fingerprinting services. It tackles Safari’s cookie-breaking privacy updates by stitching anonymous user sessions securely and statefully. The core stack runs comfortably on an 8 GB VPS using Node.js/Express, SQLite in WAL mode with tuned timeouts, Redis for rate limits and distributed locks, and an embedded ONNX vector model for semantic context matching. CPU-heavy inference is offloaded to worker threads to keep response times under 20 ms. I also implemented a soft-association algorithm that prevents accidental cart merges on shared Wi-Fi by requiring multiple overlapping sessions before linking identities. The project is fully open-source under MIT, and I’ve provided a live demo with free beta key access. I’d love feedback on the database locking strategy, worker thread design, or the privacy model. Let me know your thoughts!
Use The App To Win ₦1m
Stories are shared by community members. This article does not represent the official view of NaijaWorld — the author is solely responsible for its content.