github doobidoo/mcp-memory-service v8.23.0
v8.23.0 - Consolidation Scheduler via Code Execution API

latest releases: v10.31.2, v10.31.1, v10.31.0...
4 months ago

v8.23.0 - Consolidation Scheduler via Code Execution API

Consolidation scheduler migrated to HTTP server with Code Execution API, achieving 88% token reduction (803K tokens/year saved) and 24/7 operation independent of Claude Desktop.

Major Features

HTTP Server Integration

  • 24/7 Scheduler Operation: Consolidation runs continuously with HTTP server, independent of Claude Desktop sessions
  • APScheduler Integration: Automated scheduling with configurable time horizons (daily, weekly, monthly)
  • Persistent Operation: Scheduler survives Claude Desktop restarts and maintains state

Code Execution API

  • Ultra-Efficient Operations:
    • consolidate(): 15 tokens vs 150 tokens MCP (90% reduction)
    • scheduler_status(): 10 tokens vs 125 tokens MCP (92% reduction)
  • Compact Result Types: CompactConsolidationResult, CompactSchedulerStatus
  • Direct Python Integration: Import and call functions directly without MCP overhead

New HTTP Endpoints

  • POST /api/consolidation/trigger - Manually trigger consolidation (10-30s response)
  • GET /api/consolidation/status - Check scheduler status (<5ms response)
  • GET /api/consolidation/recommendations/{horizon} - Get consolidation recommendations (~50ms)

Token Efficiency

Massive token savings through compact API responses:

  • Before: ~900K tokens/year (MCP server retrieval overhead)
  • After: ~97K tokens/year (compact API responses)
  • Reduction: 803K tokens (88% efficiency gain)

Migration Guide

For Existing Users

  • No action required - Consolidation automatically migrates to HTTP server when enabled.

HTTP Server Setup (if not already enabled)

# Enable HTTP server in .env
export MCP_HTTP_ENABLED=true

# Restart service
systemctl --user restart mcp-memory-http.service

Backward Compatibility

  • Existing MCP consolidation tools (consolidate_memories, scheduler_status) continue working
  • All MCP tools now use Code Execution API internally for efficiency
  • No breaking changes to existing workflows

Files Modified

Code Execution API (7 files):

  • src/mcp_memory_service/api/types.py - New compact result types
  • src/mcp_memory_service/api/operations.py - Consolidation functions
  • src/mcp_memory_service/api/client.py - Client methods
  • src/mcp_memory_service/api/__init__.py - Export updates
  • src/mcp_memory_service/web/app.py - Scheduler integration
  • pyproject.toml - APScheduler dependency
  • src/mcp_memory_service/web/api/consolidation.py - New endpoint router

Documentation (3 files):

  • CHANGELOG.md - Comprehensive v8.23.0 release notes
  • README.md - Updated "Latest Release" section
  • CLAUDE.md - Updated version reference and consolidation commands

Testing Status

All features tested and verified:

  • HTTP endpoints operational
  • Scheduler initialization successful
  • Consolidation trigger tested (processed 2421 memories)
  • Background sync working correctly

What's Next

Post-release improvements planned:

  • Enhanced security for consolidation endpoints
  • Web dashboard integration for consolidation management
  • Performance optimizations for large memory sets

Full Changelog: https://github.com/doobidoo/mcp-memory-service/blob/main/CHANGELOG.md#8230---2025-

🤖 Generated with Claude Code

Co-Authored-By: Claude noreply@anthropic.com

Don't miss a new mcp-memory-service release

NewReleases is sending notifications on new releases.