Uses SQLite as the persistent storage backend for AI agent memory data, including active contexts, file indexes, spatial maps, bugs, locks, and context events.
mnehmos.synch.mcp
Global Memory Bank for AI Agents
An MCP server that provides persistent context synchronization for AI agents across sessions and projects.
Features
Active Context - Get/set current working state per project
Filing Cabinet - Index files with summaries and metadata for fast retrieval
Memory Search - Search across all indexed content
Spatial Map - "PC as Rooms" folder navigation metaphor
Bug Tracking - Log and resolve bugs for agent workflows
Lock Management - Concurrent agent access coordination
Context Events - Agent-to-agent handoff protocol
Installation
Configuration
Add to your MCP client config:
Tools
Tool | Description |
| Get current summary/focus for a project |
| Update active context state |
| Index a file with summary and metadata |
| Retrieve indexed file info |
| Search across all indexed content |
| List all projects in memory bank |
| Get folder structure as "rooms" |
| Add folder to spatial map |
| Connect two folders |
| Log a bug for later fixing |
| Get bugs by project/status |
| Mark bug as resolved |
| Lock a resource for exclusive access |
| Release a held lock |
| Check lock state |
| Emit handoff/checkpoint/error events |
| Get recent context events |
Usage
Data Storage
Data is stored in SQLite at:
Windows:
%APPDATA%\mnehmos-synch\macOS/Linux:
~/.config/mnehmos-synch/
License
MIT
Author
Mnehmos