CodeMind
Enables GitHub Copilot to understand and navigate your codebase with 20 tools for search, context, dependencies, and safety checks across projects.
README
CodeMind 🧠
Give GitHub Copilot memory across all your projects
CodeMind is an MCP server that gives GitHub Copilot 20 specialized tools for understanding your codebase.
<!-- mcp-name: io.github.MrUnreal/codemind -->
Why CodeMind?
Without it:
- ❌ Copilot creates duplicate files
- ❌ Forgets decisions you just made
- ❌ Doesn't understand your project structure
With it:
- ✅ Finds existing code before creating new files
- ✅ Remembers architectural decisions
- ✅ Understands dependencies and relationships
- ✅ Warns about breaking changes
Quick Start
Option 1: Install from PyPI (Recommended)
1. Install the package
pip install mcp-codemind
2. Configure VS Code
Add to .vscode/settings.json:
{
"mcp.servers": {
"codemind": {
"command": "python",
"args": ["-m", "codemind"]
}
}
}
3. Reload
Press Ctrl+Shift+P → "Developer: Reload Window"
Done! 🎉
Option 2: Install from Source
1. Clone and install
git clone https://github.com/MrUnreal/codemind.git
cd codemind
pip install -r requirements.txt
2. Configure VS Code
Add to your VS Code settings (Ctrl+Shift+P → "Preferences: Open User Settings (JSON)"):
{
"mcp.servers": {
"codemind": {
"command": "python",
"args": ["/full/path/to/codemind.py"],
"cwd": "${workspaceFolder}"
}
}
}
Replace /full/path/to/codemind.py with the actual path to where you cloned CodeMind.
3. Reload VS Code
Press Ctrl+Shift+P → "Developer: Reload Window"
4. Test It
Open Copilot Chat and try:
Does this project have authentication?
Done! 🎉
Need help? Check our Troubleshooting Guide
What You Get
20 AI Tools in 6 Categories:
| Category | What It Does |
|---|---|
| 🔍 Search | Find existing code, check if features exist |
| 📝 Context | Understand files, track changes, remember decisions |
| 🔗 Dependencies | See what imports what, visualize structure |
| 📊 Analysis | Code quality metrics, config auditing |
| ⚠️ Safety | Check breaking changes before refactoring |
| 🗂️ Management | Index files, find TODOs, git history |
How It Works
┌─────────────────────────────────────────────────────────────┐
│ You: "Add authentication" │
└────────────────────────┬────────────────────────────────────┘
│
v
┌─────────────────────────────────────────────────────────────┐
│ Copilot (with CodeMind): │
│ 🔍 Searching existing code... │
│ âś… Found: src/auth/jwt.py (95% match) │
│ 💡 "I found existing JWT auth. Should I extend it?" │
└─────────────────────────────────────────────────────────────┘
Without CodeMind:
❌ Creates duplicate auth files
❌ Doesn't know about existing code
❌ No context of your architecture
You ask naturally → Copilot automatically uses CodeMind → Gets smart suggestions
No explicit tool calls needed!
Example Usage
💬 "Does this project have authentication?"
✅ Found in src/auth/jwt.py
💬 "What will break if I rename UserModel?"
⚠️ 7 files will be affected
💬 "What depends on database.py?"
📋 Used by: models/, auth/, tests/
💬 "Show me TODOs"
📝 Found 12 TODOs across 5 files
Multi-Workspace Support
Work with multiple projects simultaneously:
- Each project gets its own database
- No cross-contamination
- Isolated configurations
Documentation
| Document | Purpose |
|---|---|
| Getting Started | Start here! Complete beginner's guide |
| Usage Guide | How to use with Copilot |
| Contributing | How to contribute to CodeMind |
| Tools Reference | All 20 tools explained |
| Examples | Real-world scenarios |
| FAQ | Common questions |
| Architecture | Technical details |
| Changelog | Version history |
Requirements
- Python 3.8 or higher
- VS Code with GitHub Copilot
- ~80MB for embedding model (downloaded automatically on first use)
- ~1-5MB per project for database
Status
- ✅ 20/20 tools operational
- ✅ 110+ tests passing (99%+ rate)
- ✅ Production ready
- ✅ Actively maintained
Contributing
Contributions welcome! See CONTRIBUTING.md
License
MIT - See LICENSE
Built with: Python, FastMCP, sentence-transformers, SQLite
Making Copilot smarter, one project at a time 🚀
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。