myai memory.com
myai memory.com logo

myAI Memory

Synchronizes personal memory templates across Claude interfaces, enabling consistent user preferences and response style...

Created byApr 23, 2025

myAI Memory Sync

[![smithery badge](https://smithery.ai/badge/@Jktfe/myaimemory-mcp)](https://smithery.ai/server/@Jktfe/myaimemory-mcp)
**Tired of repeating yourself to Claude every time you start a new chat?** myAI Memory Sync is a game-changing MCP tool that seamlessly synchronizes your preferences, personal details, and code standards across ALL your Claude interfaces! Just update once, and your changes instantly appear everywhere - from Claude Desktop to Claude Code, Windsurf, and Claude.ai web. With our cutting-edge caching system, memory-related queries are up to 2000x faster! Stop wasting tokens on repetitive instructions and enjoy a truly personalized AI experience.

How myAImemory-mcp Compares to Other Memory Tools

While several excellent memory tools exist for AI systems, myAImemory-mcp serves a specific purpose as a Model Context Protocol (MCP) tool:
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
**Key Advantages of myAImemory-mcp:**
  • **Privacy-First**: All data remains on your device, no personal information sent to external servers
  • **Performance**: Leverages Claude's caching capabilities for dramatic speed improvements
  • **Simplicity**: Natural language updates to your preferences across all Claude interfaces
  • **MCP Integration**: Purpose-built as a Claude MCP for seamless integration

Quick Start

Server Options

The unified server script supports multiple options:

Direct Sync Method (Simple Alternative)

For a simpler approach that doesn't require running an MCP server, you can use the unified CLI:
This script will:
  • Read from your "myAI Master.md" file
  • Update all CLAUDE.md files in your projects
  • Update your Windsurf memory settings
  • All without storing sensitive information in the git repository

Privacy and Security

  • The "myAI Master.md" file with your personal information is excluded from git tracking
  • All CLAUDE.md files are also excluded to protect your privacy
  • Use the included `.gitignore` to ensure sensitive files remain private

Supported Natural Language Commands

You can interact with myAI Memory using these natural language patterns:
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
Note: To perform a full sync across all platforms, use the command line: `node sync-memory.js`

Installation Options

Option 1: Direct Install (Recommended)

Install from npm:
Start the server:

Option 2: Run from Source

Clone and build from source:

Option 3: Docker

Build and run with Docker:

MCP Configuration

Claude Desktop Configuration

Add this to your `claude_desktop_config.json`:

Claude.ai with Smithery

  1. Visit [Smithery.ai](https://smithery.ai)
  1. Add the myAI Memory Sync MCP: ``` @Jktfe/myaimemory-mcp ```
  1. Configure with your API key in the Smithery settings

Windsurf Integration

In Windsurf, add to your `.codeium/config.json`:

HTTP Server Mode

For HTTP transport instead of stdio:

Environment Variables

Create a `.env` file with the following options:

System Prompt Integration

For best results, add this to your Claude system prompt:

Features

  • **Cross-Platform Synchronization**: Update once, syncs everywhere
  • **Lightning-Fast Recall**: Caching system with up to 2000x performance boost
  • **Natural Language Interface**: Just talk naturally to update your preferences
  • **Multiple Persona Profiles**: Switch between different presets with ease
  • **Security-Focused**: Local storage with .gitignore protection
  • **Developer-Friendly**: Full TypeScript implementation with comprehensive API

Core Architecture

myAI Memory Sync uses a modular architecture with these key components:
  • **Template Parser**: Bidirectional conversion between structured memory objects and markdown
  • **Template Storage**: Persistent storage with in-memory and file-system caching
  • **Platform Synchronizers**: Implements the `PlatformSyncer` interface for each target platform
  • **Natural Language Processor**: Extracts structured data from natural language memory commands
  • **Memory Cache Service**: Optimizes performance with multi-level caching strategies

Detailed Features

Cross-Platform Synchronization

  • **ClaudeCodeSyncer**: Updates CLAUDE.md files across all repositories
  • **WindsurfSyncer**: Manages global_rules.md in Windsurf environment
  • **ClaudeWebSyncer**: Optional Puppeteer-based synchronization with Claude.ai web interface

Intelligent Memory Management

  • **Pattern-Based Extraction**: Converts natural language to structured key-value pairs
  • **Section Detection Algorithm**: Automatically determines appropriate section for new memories
  • **Memory Template Format**: Markdown-based structure with sections, descriptions, and key-value items
  • **Context Preservation**: Updates memory sections while preserving other template content

Performance Optimization

  • **Multi-Level Caching**: In-memory caching at both template and section levels
  • **TTL-Based Cache Management**: Configurable Time-To-Live for cached content
  • **Pre-Warming**: Cache pre-population after template updates
  • **Optional Anthropic API Integration**: Accelerates memory-related queries up to 2000x

Security

  • **Local-First Architecture**: All data remains on your device
  • **Gitignore Management**: Automatically adds CLAUDE.md to .gitignore in all repositories
  • **File Permission Handling**: Fixes permissions issues for maximum compatibility
  • **Encrypted Storage**: Compatible with encrypted file systems

Memory Template Format

The system uses a structured markdown format to organize your preferences:

Technical Implementation

MemoryTemplate Schema

Platform Synchronization Interface

MCP Integration API

The myAI Memory Sync tool implements the Model Context Protocol (MCP) with the following functions:
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]

Natural Language Interface

Users can interact with the system through natural language commands:

Advanced Usage

Memory Presets

Switch between different personas easily:

Emergency Sync

When you need to fix synchronization issues across all platforms:

Command Line Interface

Development Workflow

Performance Benchmarks

Our caching system delivers incredible performance improvements:
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]
[object Object]

Security & Privacy

We take your privacy seriously:
  • All data remains locally on your device
  • CLAUDE.md files are automatically added to .gitignore
  • No data is sent to external servers (except when using the optional Anthropic API integration)
  • Works with encrypted file systems for maximum security

Troubleshooting

Common Issues

  1. **CLAUDE.md Not Updating** - Check file permissions with `ls -la CLAUDE.md` - Try emergency sync with `./emergency-sync.sh` - Verify platform paths in your `.env` file
  1. **MCP Connection Failures** - Ensure MCP server is running with `ps aux | grep myai-memory` - Check Claude Desktop logs for MCP errors - Verify your Claude Desktop configuration file
  1. **Caching Issues** - Clear cache with `node dist/cli.js --clear-cache` - Verify Anthropic API key is correctly set - Check memory file integrity with `node dist/cli.js --validate`
  1. **Natural Language Commands Not Working** - Make sure to use exactly one of the supported command patterns (see Supported Natural Language Commands section) - If Claude doesn't recognize your command, try a different pattern - For syncing across all platforms, use the direct script: `node sync-memory.js`

Manual Syncing

If you're experiencing issues with natural language commands or the MCP server:
These methods directly read from your master file and update all platforms without relying on the MCP server or natural language processing.

Logs and Debugging

Enable debug mode to see detailed logs:
Log files are stored in:
  • Linux/macOS: `~/.local/share/myai-memory/logs/`
  • Windows: `%APPDATA%\myai-memory\logs\`

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.
  1. Fork the repository
  1. Create your feature branch (`git checkout -b feature/amazing-feature`)
  1. Commit your changes (`git commit -m 'Add some amazing feature'`)
  1. Push to the branch (`git push origin feature/amazing-feature`)
  1. Open a Pull Request
We follow a standard Git workflow and CI process:
  • All PRs require passing tests and linting
  • New features should include tests
  • Major changes should update documentation
  • Follow existing code style and patterns

Documentation

For more detailed documentation, see the [Wiki](https://github.com/Jktfe/myaimemory-mcp/wiki).
API documentation is available in the `/docs` directory:

License

This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.

Contact

Project Link: [https://github.com/Jktfe/myaimemory-mcp](https://github.com/Jktfe/myaimemory-mcp)