Cursor Memory MCP Server
A persistent SQLite-backed storage server that enables Cursor IDE's AI assistant to remember information across sessions with global or project-specific scopes. It supports complete CRUD operations, tag-based organization, and keyword search to manage user preferences and context effectively.
README
Cursor Memory MCP Server
A persistent memory storage MCP (Model Context Protocol) server for Cursor IDE. Allows the AI assistant to remember information across sessions.
Features
- Store memories with global or project-specific scope
- Tag-based organization for easy categorization
- Search memories by content or tags
- CRUD operations - create, read, update, delete memories
- Context matching - update/delete by description, not just ID
- SQLite backend for performance and reliability
- Docker deployment for easy setup
Quick Start
1. Build and Run with Docker
docker-compose up -d --build
2. Add to Cursor MCP Configuration
Edit ~/.cursor/mcp.json:
{
"mcpServers": {
"cursor-memory": {
"url": "http://localhost:8081/sse",
"transport": "sse"
}
}
}
3. Restart Cursor
The memory tools will now be available to the AI assistant.
Available Tools
| Tool | Description |
|---|---|
memory_store |
Store a new memory |
memory_recall |
Retrieve all memories (call on session start) |
memory_search |
Search memories by keyword/tag |
memory_list |
List all memories with index numbers |
memory_update |
Update a memory by ID or context |
memory_delete |
Delete a memory by ID or context |
Usage Examples
Storing a Memory
Tell Cursor: "Remember that I prefer tabs over spaces"
The agent will:
- Optimize the content
- Ask for scope (global/project)
- Suggest tags
- Confirm before storing
Recalling Memories
Tell Cursor: "Check your memory" or "What do you remember?"
Searching Memories
Tell Cursor: "Check your memory for coding preferences"
Managing Memories
- "List all memories"
- "Delete memory #3"
- "Update the memory about indentation to prefer 4 spaces"
Development
Local Development (without Docker)
# Install dependencies
npm install
# Run in development mode
npm run dev
# Build for production
npm run build
npm start
Environment Variables
| Variable | Default | Description |
|---|---|---|
PORT |
8081 |
HTTP server port |
DATA_DIR |
./data |
SQLite database directory |
NODE_ENV |
development |
Environment mode |
Architecture
┌─────────────────┐ SSE ┌──────────────────────┐
│ Cursor IDE │◄────────────►│ Memory MCP Server │
│ (AI Agent) │ │ (Express + SQLite) │
└─────────────────┘ └──────────────────────┘
│
▼
┌─────────────┐
│ SQLite DB │
│ memory.db │
└─────────────┘
Port Configuration
This server runs on port 8081 by default to avoid conflicts with other MCP servers (e.g., jana on 8080).
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。