A memory management system for AI assistants to store, retrieve, and manage user information using a local database.
Simple Memory MCP is a memory management system designed for AI assistants, implementing the Model Context Protocol (MCP) to provide persistent memory storage and retrieval capabilities.
Tell your AI assistant: "Open memory" or "Open memory WEB", the AI will:
Tell your AI assistant: "Get all memories" or "Show memory list", the AI will execute the list_memory_titles
tool to retrieve all memory titles.
Tell your AI assistant: "Get 'XXX' memory content and execute", the AI will execute the get_memory_by_title
tool to retrieve the complete content of the specified memory.
Tell your AI assistant: "Help me store a memory", the AI will:
store_memory
tool to complete storage# Clone the repository
git clone https://github.com/eragonht1/simple-memory-mcp.git
# Navigate to project directory
cd simple-memory-mcp
For Windows Users:
For macOS Users:
# Option 1: Download from official website
# Visit https://nodejs.org/ and download the LTS version
# Option 2: Using Homebrew (recommended)
brew install node
For Linux Users:
# Ubuntu/Debian
sudo apt update
sudo apt install nodejs npm
# CentOS/RHEL/Fedora
sudo dnf install nodejs npm
# Arch Linux
sudo pacman -S nodejs npm
Verify Installation:
# Check Node.js version
node --version
# Check npm version
npm --version
Both commands should return version numbers (e.g., v18.17.0 for Node.js).
# Install dependencies
npm install
# Initialize database
npm run init-db
# Start MCP Server
npm start
# Start Web Interface (recommended)
node start-web.js
# or
npm run web
Use the following configuration template for any AI assistant that supports MCP:
{
"mcpServers": {
"simple-memory": {
"command": "node",
"args": ["./start-mcp.js"],
"cwd": "/path/to/your/simple-memory-mcp"
}
}
}
Configuration Steps:
/path/to/your/simple-memory-mcp
with your actual project pathNote: The start-mcp.js
script provides better error handling and user-friendly output compared to directly calling src/server.js
.
Simple Memory MCP uses a single instance design to ensure optimal resource usage and data consistency:
Multi-AI Workflow:
User: Help me store a memory
AI: Please provide the title for this memory:
User: Study Notes - MCP Protocol
AI: Please enter the memory content:
User: [Enter content]
AI: Memory "Study Notes - MCP Protocol" has been successfully stored!
Simply say any of these trigger phrases to your AI assistant:
The AI will automatically:
store_memory
- Store new memory (requires title and content)list_memory_titles
- Get list of all memory titlesget_memory_by_title
- Retrieve memory content by titledelete_memory
- Delete specified memoryopen_memory_web
- NEW! Open memory management web interface with auto browser launchFor detailed documentation, see DOCS.md which includes:
Port conflicts:
# Windows
netstat -ano | findstr :8011
taskkill /PID <PID> /F
# Linux/macOS
lsof -i :8011
kill -9 <PID>
Database issues:
rm data/memories.db
npm run init-db
Web interface auto-launch issues:
# If browser doesn't open automatically
# Check if the web server is running
curl http://localhost:8011
# Manually open the URL shown in AI response
# Example: http://localhost:8011
# Check browser availability (Linux)
which xdg-open firefox google-chrome
# Check browser availability (Windows)
where start
# Check browser availability (macOS)
which open
Simple Memory MCP 是一个专为AI助手设计的记忆管理系统,实现了模型上下文协议(MCP),为AI助手提供持久化记忆存储和检索功能。
对AI助手说:"开启记忆"或"打开记忆WEB",AI会:
对AI助手说:"获取所有记忆"或"显示记忆列表",AI会执行 list_memory_titles
工具获取所有记忆标题。
对AI助手说:"获取'XXX'记忆内容并执行",AI会执行 get_memory_by_title
工具获取指定记忆的完整内容。
对AI助手说:"帮我存储一个记忆",AI会:
store_memory
工具完成存储# 克隆仓库
git clone https://github.com/eragonht1/simple-memory-mcp.git
# 进入项目目录
cd simple-memory-mcp
Windows用户:
macOS用户:
# 方法1:从官网下载
# 访问 https://nodejs.org/ 下载LTS版本
# 方法2:使用Homebrew(推荐)
brew install node
Linux用户:
# Ubuntu/Debian
sudo apt update
sudo apt install nodejs npm
# CentOS/RHEL/Fedora
sudo dnf install nodejs npm
# Arch Linux
sudo pacman -S nodejs npm
验证安装:
# 检查Node.js版本
node --version
# 检查npm版本
npm --version
两个命令都应该返回版本号(例如:Node.js显示v18.17.0)。
# 安装依赖
npm install
# 初始化数据库
npm run init-db
# 启动MCP服务器
npm start
# 启动Web管理界面(推荐)
node start-web.js
# 或者
npm run web
以下配置模板适用于任何支持MCP的AI助手:
{
"mcpServers": {
"simple-memory": {
"command": "node",
"args": ["/path/to/your/simple-memory-mcp/start-mcp.js"],
"cwd": "/path/to/your/simple-memory-mcp"
}
}
}
配置步骤:
/path/to/your/simple-memory-mcp
替换为你的实际项目路径注意: start-mcp.js
脚本相比直接调用 src/server.js
提供了更好的错误处理和用户友好的输出信息。
Simple Memory MCP 采用单实例设计,确保最佳的资源使用和数据一致性:
多AI工作流程:
用户: 帮我存储一个记忆
AI: 请提供这个记忆的标题:
用户: 学习笔记 - MCP协议
AI: 请输入记忆的具体内容:
用户: [输入内容]
AI: 记忆 "学习笔记 - MCP协议" 已成功存储!
只需对AI助手说出以下任一触发词:
AI将自动:
store_memory
- 存储新记忆(需要标题和内容参数)list_memory_titles
- 获取所有记忆标题列表get_memory_by_title
- 根据标题检索记忆内容delete_memory
- 删除指定记忆open_memory_web
- 新功能! 打开记忆管理Web界面并自动启动浏览器详细文档请参见 DOCS.md,包含:
端口冲突:
# Windows
netstat -ano | findstr :8011
taskkill /PID <PID> /F
# Linux/macOS
lsof -i :8011
kill -9 <PID>
数据库问题:
rm data/memories.db
npm run init-db
Web界面自动启动问题:
# 如果浏览器没有自动打开
# 检查Web服务器是否运行
curl http://localhost:8011
# 手动打开AI响应中显示的URL
# 例如: http://localhost:8011
# 检查浏览器可用性 (Linux)
which xdg-open firefox google-chrome
# 检查浏览器可用性 (Windows)
where start
# 检查浏览器可用性 (macOS)
which open
MIT License
Simple Memory MCP - 让AI助手拥有持久记忆的能力 💾✨ Simple Memory MCP - Giving AI assistants persistent memory capabilities 💾✨
An Elasticsearch server focused on security and threat analysis. Requires a valid Elasticsearch license (trial, platinum, or enterprise) for connection.
Search and query patient Electronic Health Record (EHR) data using SMART on FHIR.
A read-only MCP server for MySQL, enabling LLMs to query live data using the CData JDBC Driver.
An MCP server for the Chroma embedding database, providing persistent, searchable working memory for AI-assisted development with features like automated context recall and codebase indexing.
MCP Server for OceanBase database and its tools
A read-only MCP server by CData for querying live Cloudant data with LLMs. Requires the CData JDBC Driver for Cloudant.
Update various databases (PostgreSQL, MySQL, MongoDB, SQLite) using data from CSV and Excel files.
A read-only MCP server for querying live Google Directory data using the CData JDBC Driver.
A server for storing and searching data in a VikingDB instance, configurable via command line or environment variables.
A read-only MCP server by CData for querying live Zoho Creator data using a JDBC driver.