# NiPlex MCP Server

A production-ready Model Context Protocol (MCP) server that provides AI agents with access to various tools and integrations.

## Features

This MCP server includes the following tool categories:

### GitHub Integration
- List, create, and manage repositories
- Read/write files in repositories
- Create branches, commits, issues, and pull requests
- Full repository management capabilities

### Code Execution & Sandboxes
- Daytona: Run shell commands in disposable cloud sandboxes
- E2B: Execute code with internet access in short-lived sandboxes
- Horizon: Free sandbox for quick code execution
- GUI automation via E2B desktop sandboxes

### Production SFTP Server
- List, read, write, and delete files on production servers
- Create directories and rename files
- Search for files by pattern
- Full SFTP file management

### Search & Web Scraping
- Broad internet search via You.com API or DuckDuckGo
- Extract clean text from web pages
- Cached results for efficiency

### YouTube Integration
- Search YouTube videos
- Get detailed video metadata
- Retrieve channel statistics

### Google Workspace
- Gmail: Search messages, read threads, create drafts
- Calendar: List events, create events, update events
- Drive: Search files, read/write Google Docs

### Notion Integration
- Global Search: Search workspace titles, pages, and databases
- Page Management: Retrieve, create, update, and archive pages
- Block & Content: Fetch content as Markdown, append/update/delete blocks
- Database Operations: Query, get schema, create/update databases, add items
- Collaboration: Read/write comments, fetch workspace users

### Miscellaneous
- Helper utilities for testing MCP connectivity
- List installed package versions
- View recent error logs

## Quick Start

### Prerequisites
- Python 3.8+
- Various API keys for services (optional)

### Installation

1. Clone this repository:
```bash
git clone https://github.com/Aj-Niplex/NiPlex-MCP.git
cd NiPlex-MCP
```

2. Install dependencies:
```bash
pip install -r requirements.txt
```

3. Configure environment variables (see env.example)

4. Run the server:
```bash
python server.py
```

The server will start on `http://0.0.0.0:7860` by default.

### Using with AI Agents

Add this MCP server to your AI agent configuration:
- Server URL: `http://localhost:7860` (or your deployed URL)
- API Key: Set the `MCP_API_KEY` in your agent's environment

## Tool Reference

See server.py for the complete list of available tools and their descriptions.

## Architecture

### Manager Pattern

Each integration category has its own manager class in `tools/managers/`:
- git_manager.py - GitHub operations
- sandbox_manager.py - Sandbox execution
- sftp_manager.py - SFTP file operations
- search_manager.py - Web search and scraping
- youtube_manager.py - YouTube operations
- misc_manager.py - Miscellaneous utilities
- google_manager.py - Google Workspace operations
- notion_manager.py - Notion operations

Each manager provides a `call()` method that routes tool calls to the appropriate handler.

## License

MIT License