GitHub Search MCP Server
Enables searching and exploring GitHub repositories with tools for cloning repos, searching code with regex, browsing directory structures, reading files, and extracting code outlines.
README
GitHub Search MCP Server
A Model Context Protocol (MCP) server for searching and exploring GitHub repositories. Compatible with OpenAI's ChatGPT and Responses API.
Features
-
5 MCP Tools with
readOnlyHintannotations:clone_repository- Clone repositories from a whitelisted GitHub usersearch_code- Fast grep-based code search with regex supportget_tree- Display repository directory structureread_file- Read file contents with line range supportget_outline- Extract code structure (classes, functions, methods)
-
No Authentication Required - Server handles GitHub PAT internally
-
Streamable HTTP Transport - Compatible with OpenAI's MCP integration
-
MCP Protocol 2025-06-18 - Latest specification compliance
Quick Start
1. Install Dependencies
pip install -r requirements.txt
2. Configure Environment
cp .env.example .env
# Edit .env with your GitHub PAT token
3. Run the Server
python main.py
The server will start on http://localhost:8000.
Docker
# Build
docker build -t github-mcp-server .
# Run
docker run -p 8000:8000 -e GITHUB_PAT=your_token github-mcp-server
API Endpoints
| Endpoint | Method | Description |
|---|---|---|
/ |
GET | Server info and available endpoints |
/mcp |
POST | MCP protocol endpoint (JSON-RPC) |
/health |
GET | Health check |
/capabilities |
GET | Server capabilities |
MCP Tools
clone_repository
Clone a GitHub repository from the whitelisted user.
{
"name": "clone_repository",
"arguments": {
"repo_name": "my-repo"
}
}
search_code
Search for code patterns using grep.
{
"name": "search_code",
"arguments": {
"repo_name": "my-repo",
"pattern": "def main",
"file_pattern": "*.py",
"case_sensitive": false,
"max_results": 50
}
}
get_tree
Display repository directory tree.
{
"name": "get_tree",
"arguments": {
"repo_name": "my-repo",
"path": "src",
"max_depth": 3,
"show_hidden": false
}
}
read_file
Read file contents with optional line ranges.
{
"name": "read_file",
"arguments": {
"repo_name": "my-repo",
"file_path": "src/main.py",
"start_line": 1,
"end_line": 50
}
}
get_outline
Get code outline showing classes and functions.
{
"name": "get_outline",
"arguments": {
"repo_name": "my-repo",
"file_path": "src/main.py"
}
}
Connecting to ChatGPT
- Deploy the server to a public URL (e.g., Railway, Render, or AWS)
- In ChatGPT, go to Settings > Connectors > Add Connector
- Enter your server URL (e.g.,
https://your-server.com/mcp) - Select "No authentication required"
Security
- Whitelisted User: Only repositories from the configured
ALLOWED_USERNAMEcan be cloned - Path Traversal Prevention: All file paths are validated
- No Secrets in Responses: GitHub PAT is server-side only
Environment Variables
| Variable | Description | Default |
|---|---|---|
GITHUB_PAT |
GitHub Personal Access Token | (empty) |
HOST |
Server host | 0.0.0.0 |
PORT |
Server port | 8000 |
REPO_STORAGE_PATH |
Path for cloned repos | /tmp/repos |
ALLOWED_USERNAME |
Whitelisted GitHub username | anirudhadasgupta |
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。