MCP Project Context Server

An MCP (Model Context Protocol) server that enables AI agents in any IDE to analyze and gain context about any project on your system.

Features

The server provides the following tools for AI agents:

| Tool | Description | | ----------------------- | ------------------------------------------------------ | | set_project_path | Set the path to the project you want to analyze | | get_project_structure | Get the directory tree structure of the project | | list_files | List all files in the project | | read_file | Read the content of a specific file | | read_multiple_files | Read multiple files at once | | search_in_files | Search for text/regex patterns across project files | | get_file_info | Get file metadata (size, modification date) | | get_project_summary | Get project summary (file types, technology detection) |

Installation

npm install
npm run build

Usage

1. Running the Server

npm start

Or with a project path set via environment variable:

# Windows PowerShell
$env:PROJECT_PATH="C:\path\to\your\project"; npm start

# Windows CMD
set PROJECT_PATH=C:\path\to\your\project && npm start

# Linux/Mac
PROJECT_PATH=/path/to/your/project npm start

2. IDE Configuration

Configure the MCP server in your IDE that supports the Model Context Protocol.

VS Code (with Copilot or other MCP-compatible extension)

Add to your VS Code settings or MCP configuration file:

{
  "mcpServers": {
    "project-context": {
      "command": "node",
      "args": ["/path/to/mcp-project-context/dist/mcp-server.js"],
      "env": {
        "PROJECT_PATH": "/path/to/project/you/want/to/analyze"
      }
    }
  }
}

JetBrains IDEs with GitHub Copilot (Android Studio, IntelliJ IDEA, WebStorm, PyCharm, etc.)

Open or create the mcp.json file in your JetBrains IDE (Settings → Tools → GitHub Copilot → MCP Servers, or find it in your IDE config directory) and add:

{
  "servers": {
    "project-context": {
      "type": "stdio",
      "command": "node",
      "args": ["C:\\path\\to\\mcp-project-context\\dist\\mcp-server.js"],
      "env": {
        "PROJECT_PATH": "C:\\path\\to\\project\\you\\want\\to\\analyze"
      }
    }
  }
}

Note: On Windows use double backslashes \\ or forward slashes / in paths.

Cursor, Windsurf, or other MCP-compatible editors

Check your editor's documentation for MCP server configuration. The configuration format is typically similar:

{
  "project-context": {
    "command": "node",
    "args": ["/path/to/mcp-project-context/dist/mcp-server.js"]
  }
}

Global MCP Configuration

You can also create a global configuration file:

Linux/Mac: ~/.config/mcp/servers.json
Windows: %APPDATA%\mcp\servers.json

{
  "project-context": {
    "command": "node",
    "args": ["/path/to/mcp-project-context/dist/mcp-server.js"]
  }
}

3. Using with AI Agent

Once connected, the AI agent can use these tools:

First, set the project path:

set_project_path({ projectPath: "/home/user/projects/my-app" })

Get a project overview:
```
get_project_summary()
```
Explore the structure:
```
get_project_structure({ maxDepth: 3 })
```

Read specific files:

read_file({ relativePath: "src/main.ts" })

Search in code:

search_in_files({ query: "function handleClick", filePattern: ".ts" })

Security

The server automatically ignores binary files and common directories like node_modules, .git, build
Access is restricted to files within the configured project directory
Cannot read files outside the project directory (path traversal protection)

Ignored Directories

node_modules
.git
.idea
.gradle
build
.next
dist
out
pycache
.venv, venv

Ignored File Types

Images: .jpg, .jpeg, .png, .gif, .ico, .svg
Archives: .zip, .tar, .gz
Binaries: .exe, .dll, .so, .class, .jar, .apk, .aab
Logs: .lock, .log

License

MIT

MCP Servers