mcp-engine

mcp-engine

3.9

If you are the rightful owner of mcp-engine and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcpreview.com.

MCPEngine is a production-grade, HTTP-first implementation of the Model Context Protocol (MCP), providing a secure, scalable, and modern framework for exposing data, tools, and prompts to Large Language Models (LLMs) via MCP.

MCPEngine is designed to be the 'REST for LLMs,' enabling applications to expose standardized endpoints accessible by LLMs without custom-coded integrations. It offers a robust framework for modern, cloud-native use cases, with features like built-in OAuth, HTTP-first design, and scope-based authorization. MCPEngine uses a proxy-based architecture to integrate with LLM hosts, providing seamless integration, security, scalability, and separation of concerns. The server can run anywhere, making it versatile for various deployment scenarios. MCPEngine supports tools, resources, and prompts, allowing developers to create LLM-invokable functions, provide read-only context, and use reusable conversation templates. It also supports OAuth and scope-based authorization, ensuring secure access to tools and resources.

Features

  • Built-in OAuth with Okta, Keycloak, Google SSO
  • HTTP-first design (SSE instead of just stdio)
  • Scope-based Authorization for tools, resources, and prompts
  • Seamless bridging for LLM hosts via a local proxy
  • Full backwards-compatibility with FastMCP and the official MCP SDK

Tools