Binary Cold Storage Engine

How Lexino AI achieves 70%–85% database storage reduction using native Brotli and Gzip compression.

⏱️ 3 min readPublic Document (No Sign-in Needed)

To scale database capacity while preserving full chat history, Lexino AI implements automated binary cold storage.

#How It Works

• Inactive sessions (older than 7 days) are compressed using Node.js native zlib Brotli (Quality level 6).

• The resulting binary payload is stored as a compressed buffer in PostgreSQL, reducing row footprint by up to 85%.

• Transparent Decompression: When a user re-opens an archived session, the engine auto-detects binary headers, uncompresses the message history into memory in under 2ms, and restores the chat seamlessly.

#Archiving Endpoints

• POST /api/chat/archive: Compresses inactive sessions.

• GET /api/chat/sessions: Returns all chat sessions with automatic transparent decompression.