doobidoo/mcp-memory-service v6.13.7

🐛 v6.13.7 - Critical Cloudflare Vectorize ID Fix

on GitHub

🐛 Critical Bug Fixes

Cloudflare Vectorize ID Length Issue Fixed

• Fixed critical bug: Resolved Cloudflare Vectorize vector ID length limit error
  • Root Cause: CloudflareStorage was using "mem_" + content_hash format for vector IDs (68 characters)
  • Cloudflare Limit: Vectorize has a 64-byte maximum ID length restriction
  • Solution: Removed "mem_" prefix, now using content_hash directly (64 characters)
  • Impact: Enables proper bidirectional sync between multiple machines using Cloudflare backend
  • Error Fixed: "id too long; max is 64 bytes, got 68 bytes" when storing memories

🔄 Multi-Machine Sync Capability

New Architecture

• Cloudflare as Central Hub: D1 Database + Vectorize Index + Workers AI
• Bidirectional Sync: Real-time memory synchronization across machines
• Backup Strategy: Local sqlite_vec files maintained for fallback
• Global Access: Fast access from anywhere via Cloudflare CDN

Use Cases

• ✅ Replace failed narrowbox/central servers
• ✅ Sync memories across development machines
• ✅ Team collaboration with shared memory context
• ✅ Robust disaster recovery strategies

⚠️ Breaking Changes

Vector ID Format Change

• Before: "mem_" + content_hash (68 characters)
• After: content_hash only (64 characters)
• Migration: Existing Cloudflare users may need to re-import memories

📚 Documentation Updates

New Documentation

• Multi-Machine Sync Wiki Page: Comprehensive implementation guide
• Updated Cloudflare Setup: Added bidirectional sync section

🎯 Real-World Success Story

Failed Server Recovery

Scenario: Development team's narrowbox server failed due to power issues

• Challenge: 978 memories on M2 MacBook, old MacBook Pro needed access
• Solution: Migrated to Cloudflare with v6.13.7 ID fix
• Result: 1062 memories syncing perfectly across both machines
• Time: 15-minute implementation
• Reliability: More robust than previous single-server setup

📦 Installation

# Upgrade existing installation
pip install --upgrade mcp-memory-service

# Verify version
memory --version
# Should show: MCP Memory Service v6.13.7

🚀 Quick Setup for Multi-Machine Sync

1. Configure Cloudflare

export CLOUDFLARE_API_TOKEN="your-token"
export CLOUDFLARE_ACCOUNT_ID="your-account"
export CLOUDFLARE_D1_DATABASE_ID="your-d1-id"
export CLOUDFLARE_VECTORIZE_INDEX="mcp-memory-index"
export MCP_MEMORY_STORAGE_BACKEND="cloudflare"

2. Update Claude Desktop Config

{
  "mcpServers": {
    "memory": {
      "command": "/path/to/memory",
      "args": ["server"],
      "env": {
        "MCP_MEMORY_STORAGE_BACKEND": "cloudflare",
        "CLOUDFLARE_API_TOKEN": "your-token",
        "CLOUDFLARE_ACCOUNT_ID": "your-account",
        "CLOUDFLARE_D1_DATABASE_ID": "your-d1-id",
        "CLOUDFLARE_VECTORIZE_INDEX": "mcp-memory-index"
      }
    }
  }
}

This release enables robust, distributed memory management for the first time in MCP Memory Service history! 🎉