mcp-framework

Overview

MCP Framework is a production-ready Rust implementation for building AI agents with first-class MCP support. It provides both MCP server and client functionality with multi-LLM support including Claude and OpenAI integrations. The framework features a web-based inspector for testing tools, browser automation capabilities via Playwright, and built-in session management. It's designed for performance with zero-cost abstractions and memory safety guarantees.

Try asking AI

After installing, here are 6 things you can ask your AI assistant:

• echo

  String echo utility

  Try:Print hello worldTry:Echo the message 'test'
• calculator

  Math: add, subtract, multiply, divide, power, sqrt

  Try:What is 15 + 27?Try:Calculate the square root of 64
• get_weather

  Weather lookup for cities worldwide

  Try:What's the weather in Tokyo?Try:Check if it's raining in London
• search_text

  Find pattern occurrences in text

  Try:Find all occurrences of 'error' in this textTry:Search for the word 'important'
• string_length

  Get character count

  Try:How many characters are in 'hello world'?Try:Count the letters in this string
• text_reverse

  Reverse text strings

  Try:Reverse the string 'hello'Try:Mirror this text
• json_parser

  Validate and format JSON

  Try:Validate this JSON: {"key": "value"}Try:Format this JSON string properly
• http_status

  Look up HTTP status codes

  Try:What does HTTP 404 mean?Try:Explain status code 200
• greet

  Greet someone

  Try:Say hello to AliceTry:Greet the user

Comparable tools

Installation

# Requires Rust 1.70+
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
git clone https://github.com/koki7o/mcp-framework
cd mcp-framework
cp .env.example .env
# Edit .env and add API keys

# Run examples
cargo run --example server_with_tools
cargo run --example anthropic_agent_demo_with_tools
cargo run --example openai_agent_demo_with_tools

{
  "mcpServers": {
    "mcp-framework": {
      "command": "cargo",
      "args": ["run", "--example", "server_with_tools"],
      "env": {
        "ANTHROPIC_API_KEY": "your-api-key"
      }
    }
  }
}

FAQ

What makes this framework different from other MCP implementations?

This is a production-ready Rust framework with first-class MCP support, built-in web inspector, and multi-LLM integration. It's designed for performance and memory safety.

Which LLM providers are supported?

Currently supports Claude and OpenAI with plans to add Gemini, Groq, DeepSeek, and Mistral. The framework is designed to easily accommodate additional providers.

How do I create custom MCP tools?

Implement the ToolHandler trait and register your tools with the McpServer. The framework provides a clear API and examples to get started quickly.

On Hacker News

Recent discussion from the developer community.

• Show HN: MCP-framework – Build MCP servers and AI agents in Rust▲ 1
  Story by wiwoworld · 2025-11-11

Compare mcp-framework with