doobidoo/mcp-memory-service v8.39.0

v8.39.0 - Storage-Layer Date-Range Filtering Performance Optimization

on GitHub

🚀 Performance Optimization Release

This release delivers a 10x performance improvement for analytics queries by moving date-range filtering from the application layer to the database layer.

🎯 Key Highlights

• ⚡ 10x Faster Analytics - Moved date-range filtering to database layer (#238)
• 📉 97% Data Transfer Reduction - 50MB → 1.5MB for 10,000 memories
• 🚀 Dramatic Response Time Improvements:
  • SQLite-vec: 500ms → 50ms (10x improvement)
  • Cloudflare: 2-3s → 200ms (10-15x improvement)
• 🎯 Scalability - Efficiently handles databases with >10,000 memories
• 🔧 Database-Level Filtering - Leverages indexes on created_at field

📋 What Changed

• Added get_memories_by_time_range() to Cloudflare backend with D1 database filtering
• Updated memory growth endpoint to use database-layer queries instead of fetching all memories
• Database optimization - Pushes filtering to WHERE clauses, leverages indexed columns

🔗 Related

• Resolves: Issue #238
• PR: #252
• Commits: faf9b8d, 8e38471

📖 Documentation

• CHANGELOG
• Issue #238
• PR #252

Full Changelog: v8.38.1...v8.39.0