One MCP, developed by Burugo, is an MCP server tool built with Go for its backend and React for its web interface. It functions as a centralized reverse-proxy platform, designed to manage and connect various Model Context Protocol (MCP) services from different providers. This free platform helps enhance interaction between Large Language Models (LLMs) and applications by providing structured context management. It enables AI agents to read live data and run commands across external tools, making it a key component for integrating AI with diverse services.

What does One MCP connect to?

One MCP connects to and manages various Model Context Protocol (MCP) services from different providers. It acts as a centralized reverse-proxy platform for these services. The Model Context Protocol itself is a standardized method for enhancing interaction between Large Language Models (LLMs) and applications. It provides structured context management, allowing AI agents to read live data and run commands across external tools. This platform supports stdio, SSE (Server-Sent Events), and streamable HTTP transport protocols for managing these connected MCP services.

Core features and tools

The tool's primary function is to manage and group other MCP services. It offers Service Management to install, configure, and monitor MCP services. It also provides Service Groups, which combine multiple MCP services into a single endpoint. These Service Groups are then exposed as "Anthropic Skills for Claude Code & Droid," meaning the callable tools for AI agents are the aggregated MCP services that One MCP presents. It does not list its own specific callable tools in the same manner as a direct tool.

Who is it for?

One MCP is for developers, teams, or organizations needing to manage multiple MCP services and integrate them with AI clients. It supports workflows like centralizing management of diverse MCP services and creating AI skills from service groups. The platform also provides analytics to track usage, request rates, response times, and system health of managed services. This is not the right fit if you only need a single, standalone MCP service without the need for aggregation or centralized management.

How to get started

Access is granted through a standard OAuth authorization flow, integrating with GitHub and Google login. Initial setup involves accessing the application and using a default login before configuring OAuth.

Use cases

• Centralized Management of MCP Services: Discover, install, configure, and monitor various MCP services from different providers through a single, unified platform.
• Creation of AI Skills from Service Groups: Combine multiple MCP services into a single endpoint and export them as Anthropic Skills for AI clients, enabling AI agents to access a curated and aggregated set of tools.
• Real-time Analytics and Monitoring: Track usage, request rates, response times, and system health of managed MCP services.

One MCP provides a valuable platform for managing and aggregating Model Context Protocol services, particularly for teams working with AI agents. Its ability to group services into Anthropic Skills for clients like Claude Code & Droid simplifies AI integration. The multi-user access with role-based control via OAuth adds to its utility for organizational use. Try combining multiple MCP services into a Service Group and exporting them as an Anthropic Skill.