mcp-tool-kit

Agentic abstraction layer for building high precision vertical AI agents written in python for Model Context Protocol.

101 stars PythonAI & Machine Learning Updated Oct 23, 2025

agentic-aienterprisegenerative-aillmmcpmcp-servermodelcontextprotocolpython

Documentation

MCP Tool Kit

A comprehensive toolkit for the Model Context Protocol (MCP) with automatic tool discovery and multiple transport options.

📚 **Official Documentation | 🌐 Website | 💬 Community**

🚀 **Built by UsefulAI | 📖 From GitHub to Commercial Success**

🚀 Quick Install - No Git Required!

Choose your platform and run ONE command:

🖥️ Windows (PowerShell)

powershell

irm https://raw.githubusercontent.com/getfounded/mcp-tool-kit/main/install.ps1 | iex

🍎 macOS (Terminal)

bash

bash 
Manual Download

1. Download the latest release:
   - [Windows: MCP-Tool-Kit-Windows.zip](https://github.com/getfounded/mcp-tool-kit/releases/latest)
   - [Mac/Linux: MCP-Tool-Kit-Unix.tar.gz](https://github.com/getfounded/mcp-tool-kit/releases/latest)

2. Extract and run:
   - **Windows**: Run `install.bat`
   - **Mac/Linux**: Run `./install.sh`

Git Installation (for developers)

1. **Clone the repository**

git clone https://github.com/getfounded/mcp-tool-kit.git

cd mcp-tool-kit

code

2. **Run the launcher**
   - **Windows**: Double-click `launch.bat`
   - **Mac/Linux**: Run `./launch.sh`

Docker-Only Installation

docker run -it --name mcp-toolkit ghcr.io/getfounded/mcp-tool-kit:latest

code

## 🔧 Features

### Dynamic Tool Registration
Tools are automatically discovered and registered at runtime - no manual configuration needed!

### Multiple Transport Options
- **stdio**: For Claude Desktop integration
- **SSE (Server-Sent Events)**: For web-based access

### 120+ Available Tools
- 📁 **File System**: Read, write, and manage files
- 🕐 **Time Tools**: Timezone conversions and time operations
- 🌐 **Web Search**: Brave Search integration
- 🤖 **Browser Automation**: Playwright-based browser control
- 📊 **Data Analysis**: Yahoo Finance, FRED, World Bank data
- 📄 **Document Tools**: PDF, Excel, PowerPoint manipulation
- 🎯 **And many more!**

### Easy Configuration
Control everything through a simple `config.yaml` file:

enabled_tools:

filesystem: true

time_tools: true

brave_search: true

# ... more tools

tool_config:

filesystem:

allowed_directories: ["~/Documents", "~/Downloads"]

brave_search:

max_results: 10

code

## 📋 Environment Setup

1. Copy `.env.template` to `.env`
2. Add your API keys:

BRAVE_SEARCH_API_KEY=your_key_here

NEWS_API_KEY=your_key_here

FRED_API_KEY=your_key_here

# ... other keys

code

## 💾 Storage Setup

MCP Tool Kit uses a dedicated storage directory for file operations:

- **Windows**: `C:\Users\{Username}\mcp-tool-kit-storage`
- **Mac/Linux**: `~/mcp-tool-kit-storage`

### Automatic Setup
The storage directory is automatically created when you first run the Docker container. You can also manually set it up:

**Windows (PowerShell)**:

.\scripts\setup-storage.ps1

code

**Mac/Linux**:

./scripts/setup-storage.sh

code

### Storage Structure

mcp-tool-kit-storage/

├── documents/ # For document files

├── downloads/ # For downloaded files

└── workspace/ # For temporary work files

code

All file operations through the filesystem tool will be restricted to this directory for security.

## 🔌 Connecting to Claude Desktop

1. Start the server in stdio mode using the launcher
2. Add to Claude Desktop configuration:

{

"mcpServers": {

"mcp-tool-kit": {

"command": "docker",

"args": ["exec", "-i", "mcp-server", "python", "mcp_server_v2.py"]

}

code

## 🌐 Using SSE Mode

1. Start the server in SSE mode using the launcher
2. Access the server at `http://localhost:8080`
3. Use the SSE endpoints:
   - `/sse` - Server-Sent Events stream
   - `/messages` - Send messages to the server

## 🛠️ Creating Custom Tools

Tools now use a standardized base class system:

from app.tools.base_tool import BaseTool

class MyCustomTool(BaseTool):

def get_name(self) -> str:

return "My Custom Tool"

def get_tools(self) -> Dict[str, Callable]:

return {

"my_function": self.my_function

}

async def my_function(self, param: str, ctx: Context = None) -> str:

return f"Processed: {param}"

code

See the [developer documentation](https://docs.mcp-tool-kit.com) for detailed guides.

## 📚 Documentation

Full documentation is available at [https://docs.mcp-tool-kit.com](https://docs.mcp-tool-kit.com).

To run documentation locally:
1. `cd docs`
2. `npm install`
3. `npm start`

## 🐳 Docker Commands

### Using the Launcher (Recommended)
The launcher scripts handle all Docker operations automatically.

### Manual Commands

Start server

docker-compose up -d

View logs

docker-compose logs -f

Stop server

docker-compose down

Rebuild after changes

docker-compose build --no-cache

code

## 🤝 Contributing

1. Fork the repository
2. Create your feature branch
3. Add your tool to `app/tools/`
4. Submit a pull request

## 📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

## 🆘 Support

- [GitHub Issues](https://github.com/getfounded/mcp-tool-kit/issues)
- [Documentation](https://docs.mcp-tool-kit.com)
- [Discord Community](#) (coming soon)

Frequently asked questions

What is mcp-tool-kit?

mcp-tool-kit is Agentic abstraction layer for building high precision vertical AI agents written in python for Model Context Protocol.

How do I install mcp-tool-kit?

Open the GitHub repository and follow its README. Most MCP servers are added to your client's MCP config, then called by your agent.

Is mcp-tool-kit open source?

Yes — it is hosted on GitHub at https://github.com/getfounded/mcp-tool-kit and has 101 stars.

Related MCP tools

web-agent-protocol

🌐Web Agent Protocol (WAP) - Record and replay user interactions in the browser with MCP support Python-based implementation.

479 Python

ai-agentsai-toolsbrowser-automation+11

fastapi_mcp

Expose your FastAPI endpoints as Model Context Protocol (MCP) tools, with Auth! Python-based implementation. Trusted by 11000+ developers.

11,002 Python

aiauthenticationauthorization+10

mcp-client-for-ollama

A text-based user interface (TUI) client for interacting with MCP servers using Ollama. Features include multi-server, dynamic model switching, streaming res...

351 Python

aicommand-line-toolgenerative-ai+17

gpt-researcher

An LLM agent that conducts deep research (local and web) on any given topic and generates a long report with citations. Built for the Model Context Protocol to

24,026 Python

agentaiautomation+8

maxkb

🔥 MaxKB is an open-source platform for building enterprise-grade agents. MaxKB 是强大易用的开源企业级智能体平台。 for the Model Context Protocol. Enhance AI assistants with po

19,062 Python

agentagentic-aichatbot+11

mcp-agent

Build effective agents using Model Context Protocol and simple workflow patterns Python-based implementation. Trusted by 7600+ developers.

7,669 Python

agentsaiai-agents+5

Run your own MCP server? See who uses it and what to fix.

Measure it with TrackMCP