Making MCPs Accessible for Everyone 🚀

<div align="center"> <picture> <img src="https://raw.githubusercontent.com/klavis-ai/klavis/main/static/klavis-ai.png" width="80"> </picture> </div>

<h1 align="center">Making MCPs Accessible for Everyone 🚀</h1>

📚 TL;DR

Klavis AI is building open-source infrastructure to make Model Context Protocols (MCPs) easy for everyone. We provide:

• 💬 Slack & Discord Clients: Run MCPs directly from your favorite messaging platforms
• ☁️ Hosted MCP Servers: Access powerful tools without infrastructure management
• 🎛️ Simple Web UI: Configure and manage everything with no coding required

Whether you're a non-technical user wanting to leverage AI workflows or a developer looking to build and scale MCPs, Klavis makes it simple.

🎬 Watch Our Demo

See how easy it is to use MCPs (like our Report Generator, YouTube tools, and Document Converters) directly from Slack/Discord:

You can find more case study videos on our website.

🚀 Quick Start

For detailed MCP client setup instructions, please refer to the platform-specific documentation:

• Discord Bot Setup Guide - Complete instructions for setting up and running the Discord bot
• Slack Bot Setup Guide - Step-by-step guide for creating a Slack app and connecting it

For detailed MCP server setup instructions, please refer to the README file under each server folder.

• Discord: For Discord API integration
• Document Conversion (Pandoc): Convert between various file formats using Pandoc
• Firecrawl: For web crawling and data collection
• Firecrawl Deep Research: For deeper web research tasks
• GitHub: Perform GitHub repository operations
• Markitdown: Markdown conversion tool
• Postgres: For PostgreSQL database operations
• Report Generation: Create professional reports from various data sources
• Resend: For email services
• Slack: For Slack API integration
• Supabase: For database operations
• YouTube: Download, analyze, and transform YouTube content

🏗️ Architecture

Klavis consists of two main components:

MCP Servers

Located in the mcp_servers/ directory, these service-specific modules expose capabilities as tools:

• Report Generation: Create professional reports from various data sources
• YouTube: Download, analyze, and transform YouTube content
• Document Conversion: Convert between various file formats using Pandoc
• GitHub: Perform GitHub repository operations
• Slack: For Slack API integration
• Supabase: For database operations
• Firecrawl: For web crawling and data collection
• Resend: For email services
• Postgres: For PostgreSQL database operations
• Discord: For Discord API integration

MCP Clients

Located in the mcp_clients/ directory, these client applications connect to MCP servers and interface with end-user platforms:

• Discord Bot: Interactive AI assistant for Discord
• Slack Bot: Interactive AI assistant for Slack
• Base Client: Shared functionality for all platform clients

🧩 Extending Klavis

Adding a New Tool

1. Create a new directory in mcp_servers/
2. Implement the MCP server interface
3. Register your tools with appropriate schemas
4. Connect to your client through the standard SSE protocol

Adding a New Client Platform

1. Create a new client module in mcp_clients/
2. Extend the base_bot.py functionality
3. Implement platform-specific message handling
4. Connect to MCP servers using mcp_client.py

🤝 Contributing

We love contributions! Join our Discord community to discuss ideas and get help.

📚 Citation

If you use Klavis in your research or project, please cite:

@software{klavis2024,
  author = {Klavis AI},
  title = {Klavis: Open-Source Infrastructure for Model Context Protocols},
  year = {2024},
  publisher = {GitHub},
  url = {https://github.com/klavis-ai/klavis}
}

📜 License

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

<div align="center"> 🙏 Thanks for checking out Klavis AI! We're excited to hear your thoughts and build this with the community. </div>