Command Palette

Search for a command to run...

UnylyUnyly
Весь каталог

Like I Said Memory

БесплатноНе проверен

Like I Said Memory — Model Context Protocol server

GitHubEmbed

Описание

Like I Said Memory — Model Context Protocol server

README

A powerful Model Context Protocol (MCP) server that provides persistent memory capabilities for AI assistants like Claude Desktop, Cursor, and Windsurf, with an elegant web dashboard for memory management.

Like_I_Said_Cover

✨ Features

🔧 MCP Server Capabilities

  • Persistent Memory Storage - Add, retrieve, update, and delete memories
  • Context-Aware Storage - Store memories with rich context and metadata
  • Multi-Client Support - Works with Claude Desktop, Cursor, and Windsurf
  • JSON-Based Storage - Simple, readable memory.json file
  • Safe Configuration - Preserves existing MCP server configurations

🎨 Web Dashboard

  • Modern React Interface - Beautiful dark theme with violet accents
  • Real-Time Statistics - View memory counts, recent additions, and analytics
  • Advanced Search & Filtering - Find memories by content, tags, or metadata
  • Tag-Based Organization - Organize memories with custom tags
  • Full CRUD Operations - Create, read, update, and delete memories via UI
  • Mobile Responsive - Works on desktop and mobile devices

🚀 Installation & Updates

  • One-Click Installation - Single batch file installs everything
  • Smart Configuration - Automatically detects and configures AI assistants
  • Safe Updates - Update system preserves all data and configurations
  • Cross-Platform - Works on Windows with Node.js

📦 Quick Installation

⚠️ IMPORTANT: Windows Path Requirements

For Windows users (especially Cursor): The installation path MUST NOT contain spaces. Cursor has issues with folder paths containing spaces.

✅ Good paths:

  • C:\MCP\like-i-said-mcp-server
  • D:\Development\like-i-said-mcp-server
  • C:\Users\YourName\mcp-servers\like-i-said-mcp-server

❌ Bad paths (will cause issues):

  • D:\MY PROJECTS\AI\LLM\... (contains spaces)
  • C:\Program Files\... (contains spaces)
  • C:\Users\Your Name\... (contains spaces)

Option 1: One-Click Installer (Recommended)

Complete setup including dashboard:

  1. Create a folder without spaces (e.g., C:\MCP\)
  2. Download install-mcp-memory-server.bat
  3. Place it in your folder (e.g., C:\MCP\install-mcp-memory-server.bat)
  4. Run it:
install-mcp-memory-server.bat

The installer will:

  • ✅ Download and install the memory server
  • ✅ Configure Claude Desktop, Cursor, and/or Windsurf (your choice)
  • ✅ Set up the web dashboard
  • ✅ Create your personal memory database
  • ✅ Preserve any existing MCP configurations

Option 2: Manual Installation

If you prefer manual control or the installer doesn't work:

Step 1: Download and Setup

  1. Create a folder without spaces (e.g., C:\MCP\)
  2. Clone or download the repository:
cd C:\MCP
git clone https://github.com/endlessblink/Like-I-Said-memory-mcp-server.git
cd Like-I-Said-memory-mcp-server
npm install

Step 2: Manual Configuration

For Claude Desktop:

  1. Open: %APPDATA%\Claude\claude_desktop_config.json
  2. Add this to the mcpServers section:
{
  "mcpServers": {
    "like-i-said-memory": {
      "command": "node",
      "args": ["C:\\MCP\\Like-I-Said-memory-mcp-server\\server.js"]
    }
  }
}

For Cursor:

  1. Open: %USERPROFILE%\.cursor\mcp.json
  2. Add this to the mcpServers section:
{
  "mcpServers": {
    "like-i-said-memory": {
      "command": "node",
      "args": ["C:\\MCP\\Like-I-Said-memory-mcp-server\\server.js"]
    }
  }
}

For Windsurf:

  1. Open: %USERPROFILE%\.codeium\windsurf\mcp_config.json
  2. Add this to the mcpServers section:
{
  "mcpServers": {
    "like-i-said-memory": {
      "command": "node",
      "args": ["C:\\MCP\\Like-I-Said-memory-mcp-server\\server.js"]
    }
  }
}

Important: Replace C:\\MCP\\Like-I-Said-memory-mcp-server\\server.js with your actual installation path (use double backslashes in JSON).

Step 3: Test Installation

  1. Restart your AI assistant completely
  2. Test with: Can you store "manual installation successful" in memory?

📖 For detailed manual installation instructions, see: INSTALLATION-GUIDE.md

🎯 Usage

MCP Memory Functions

After installation, these functions are available in your AI assistants:

// Store a memory
add_memory("project_idea", "Build a personal finance app with React", {
  "tags": ["project", "react", "finance"],
  "priority": "high"
})

// Retrieve a memory
get_memory("project_idea")

// List memories with prefix
list_memories("project_")

// Delete a memory
delete_memory("project_idea")

Web Dashboard

Start the dashboard to manage memories visually:

cd your-installation-folder
npm run dev:full

Then open: http://localhost:5173

Dashboard Features:

  • 📊 Dashboard Tab - Statistics and recent memories overview
  • 📝 Memories Tab - Full table with search, filter, and management
  • Add Memories - Create new memories with tags and context
  • ✏️ Edit/Delete - Modify or remove existing memories
  • 🔍 Search & Filter - Find memories quickly by content or tags

🔄 Updates

For Existing Users

Update to get new features while preserving all your data:

# Download and run the updater:
update-mcp-memory-server.bat

The updater will:

  • 🔍 Find your installation automatically
  • 💾 Backup your data safely
  • ⬇️ Download the latest version
  • 🔄 Preserve all memories and configurations
  • ✅ Install new features and improvements

Available Update Commands

# Check current version
npm run check-updates

# Update everything
npm run update

# Stop running servers
npm run kill-servers

📁 Project Structure

like-i-said-mcp-server/
├── 📄 server.js                     # Main MCP server
├── 📄 dashboard-server.js           # Web dashboard API
├── 📄 memory.json                   # Your memory storage
├── 📄 package.json                  # Dependencies and scripts
├── 🗂️ src/                          # React dashboard source
│   ├── 📄 App.tsx                   # Main dashboard component
│   └── 🗂️ components/              # UI components
├── 🗂️ public/                       # Static assets
├── 📄 install-mcp-memory-server.bat # One-click installer
├── 📄 update-mcp-memory-server.bat  # Smart updater
└── 📄 README.md                     # This file

⚙️ Configuration

The installer automatically configures your AI assistants:

Claude Desktop

Location: %APPDATA%\Claude\claude_desktop_config.json

Cursor

Location: %USERPROFILE%\.cursor\mcp.json

Windsurf

Location: %USERPROFILE%\.codeium\windsurf\mcp_config.json

All configurations add the like-i-said-memory server alongside your existing MCP servers.

🛠️ Development

Start Development Environment

# Start both API server and React dev server
npm run dev:full

# Or start individually:
npm run dashboard  # API server (port 3001)
npm run dev        # React dev server (port 5173)

Available Scripts

npm start           # Start MCP server only
npm run dashboard   # Start dashboard API server
npm run dev         # Start React development server
npm run dev:full    # Start both dashboard and React dev server
npm run build       # Build dashboard for production
npm run preview     # Preview production build

🔧 Troubleshooting

Common Issues

"No tools available" or "Failed to create client" in Cursor:

  • 🚨 MOST COMMON ISSUE: Path contains spaces! Move to C:\MCP\ or similar path without spaces
  • Restart Cursor completely - Close and reopen the application
  • Verify server path - Ensure the path in mcp.json is correct with double backslashes
  • Test server manually: node C:\MCP\Like-I-Said-memory-mcp-server\server.js
  • Check JSON syntax - Ensure no trailing commas or syntax errors

"Port already in use" error:

npm run kill-servers

"Installation not found" during update:

  • Run the updater from your installation directory
  • Or download fresh installer: install-mcp-memory-server.bat

MCP functions not available:

  • Restart your AI assistant (Claude Desktop, Cursor, or Windsurf)
  • Check that the server path in configs is correct
  • Verify the MCP server is running: npm start

Dashboard won't load:

  • Make sure both servers are running: npm run dev:full
  • Check for port conflicts: npm run kill-servers
  • Verify Node.js and npm are installed

Path with Spaces Issues (Windows):

# If you have spaces in your path, move the installation:
# From: D:\MY PROJECTS\AI\... 
# To:   C:\MCP\like-i-said-mcp-server\

# Then update your config files with the new path

Getting Help

  • Check the troubleshooting section in UPDATE-GUIDE.md
  • Create an issue on GitHub with your error message
  • Include your operating system and Node.js version

📋 Requirements

  • Windows (tested on Windows 10/11)
  • Node.js v14 or higher
  • Git (for installation and updates)
  • Claude Desktop and/or Cursor and/or Windsurf

🔐 Security & Privacy

  • Local Storage - All memories stored locally in memory.json
  • No Cloud Sync - Data never leaves your machine
  • Config Safety - Installer preserves existing MCP configurations
  • Open Source - Full source code available for review

📄 License

MIT License - see LICENSE file for details.

🤝 Contributing

Contributions welcome! Please:

  1. Fork the repository
  2. Create a feature branch
  3. Make your changes
  4. Test thoroughly
  5. Submit a pull request

🎉 Changelog

v2.0.0 (Latest)

  • 🎨 New React Dashboard - Modern UI with dark theme
  • 🔍 Advanced Search - Filter memories by content, tags, context
  • 🏷️ Tag System - Organize memories with custom tags
  • 📊 Statistics - View memory analytics and recent activity
  • 🛡️ Safe Updates - Smart updater preserves all data
  • Performance - Faster memory operations and UI
  • 🔧 Better Config - Improved MCP configuration handling

v1.0.0

  • 🚀 Initial Release - Basic MCP server with memory functions
  • 📦 Auto-Installer - One-click setup for all AI assistants
  • 💾 Persistent Storage - JSON-based memory system

Made with ❤️ for the AI community

from github.com/Tamirido30/Like-I-Said-memory-mcp-server

Установка Like I Said Memory

У этого сервера нет опубликованного пакета — он собирается из исходников. Открой репозиторий и следуй инструкции в README.

▸ github.com/Tamirido30/Like-I-Said-memory-mcp-server

FAQ

Like I Said Memory MCP бесплатный?

Да, Like I Said Memory MCP бесплатный — установка в пару кликов через Unyly без оплаты.

Нужен ли API-ключ для Like I Said Memory?

Нет, Like I Said Memory работает без API-ключей и переменных окружения.

Like I Said Memory — hosted или self-hosted?

Self-hosted: сервер запускается локально на твоей машине командой из раздела установки.

Как установить Like I Said Memory в Claude Desktop, Claude Code или Cursor?

Открой Like I Said Memory на unyly.org, выбери вкладку своего клиента (Claude Desktop, Claude Code, Cursor) и нажми Install — конфиг сгенерируется автоматически, без правки JSON.

Похожие MCP

Compare Like I Said Memory with

Не уверен что выбрать?

Найди свой стек за 60 секунд

Автор?

Embed-бейдж для README

Похожее

Все в категории development