Codebase Explorer MCP Server
Analyzes any codebase to detect project type, languages, frameworks, entry points, API routes, controllers, services, components, tests, and more.
README
codebase-explorer-mcp
An MCP server that analyzes any codebase — detects project type, languages, framework, entry points, API routes, controllers, services, components, tests, and more.
Quick start
1. Install uv
# Windows
powershell -c "irm https://astral.sh/uv/install.ps1 | iex"
# macOS / Linux
curl -LsSf https://astral.sh/uv/install.sh | sh
Verify: uv --version
Why
uv?uv.lockpins exact dependency versions (likepackage-lock.json), so everyone gets identical environments.mcp devalso usesuvxunder the hood.
2. Setup the project
cd /path/to/codebase-explorer-mcp
uv sync
This reads uv.lock, creates .venv, and installs all dependencies.
3a. Use with Claude Desktop
Find your venv Python path:
# From the project folder:
.venv\Scripts\python.exe --version
Add to claude_desktop_config.json (%APPDATA%\Claude\claude_desktop_config.json):
{
"mcpServers": {
"codebase-explorer": {
"command": "C:\\full\\path\\to\\codebase-explorer-mcp\\.venv\\Scripts\\python.exe",
"args": ["server.py"],
"cwd": "C:\\full\\path\\to\\codebase-explorer-mcp"
}
}
}
Restart Claude Desktop.
3b. Test with MCP Inspector (development)
cd /path/to/codebase-explorer-mcp
mcp dev server.py
Opens http://localhost:6274 — test all tools live in a web UI.
How to use
You only need to provide the repo path — the server auto-discovers everything else.
| Tool | What you give it | What it returns |
|---|---|---|
analyze_structure |
path |
Languages, file count, entry points, directory tree |
find_features |
path |
Routes, controllers, components, tests, interfaces |
get_feature_detail |
path, name |
Code context around a feature (auto-locates file + line) |
generate_architecture |
path |
Layered Mermaid diagram — entry points, API, services, data, UI with import connections |
analyze_dead_code |
path |
Functions, classes, and exports defined but never referenced elsewhere |
get_git_info |
path |
Git branch, recent commits, contributors, hot files, uncommitted changes |
find_secrets |
path |
Hardcoded API keys, tokens, passwords, private keys, database credentials |
search_symbols |
path, pattern |
Regex search results across all source files |
list_project_types |
(none) | List of project types and frameworks the server can detect |
Example flow
- Get an overview:
analyze_structure(path="C:\\MyProject") - See the big picture:
generate_architecture(path="C:\\MyProject")→ Renders a Mermaid diagram with layers: Entry → API → Logic → Data → UI, plus import graph edges - Check git health:
get_git_info(path="C:\\MyProject")→ Branch, recent commits, contributors, hot files, uncommitted changes - Scan for leaks:
find_secrets(path="C:\\MyProject")→ Hardcoded API keys, tokens, passwords, and database credentials - Find dead code:
analyze_dead_code(path="C:\\MyProject")→ Lists functions, classes, and exports defined but never imported or referenced elsewhere - Find specific features:
find_features(path="C:\\MyProject")→ See all API routes, controllers, tests listed with their file locations - Inspect one:
get_feature_detail(path="C:\\MyProject", name="get_users")→ Shows code context around that feature, no need to know the file or line number - Search:
search_symbols(path="C:\\MyProject", pattern="async def")→ Find all async functions in the codebase
Alternative setup (without uv)
python -m venv venv
venv\Scripts\activate # Windows
# source venv/bin/activate # macOS / Linux
pip install "mcp[cli]>=1.6.0"
python server.py
Supported Languages
Python, JavaScript/TypeScript, Java, Go, Ruby, C#, Rust, PHP, Kotlin, Swift, and more.
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。
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。
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。
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。
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。
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。
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。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。