agentor

Overview

Agentor is an open-source platform that provides a complete solution for developing AI agents with MCP (Model Context Protocol) integration. It offers LiteMCP, a FastAPI-compatible MCP server implementation with decorator API, allowing developers to build custom MCP servers easily. The framework supports agent-to-agent communication through its A2A protocol, agent skills for specialized tasks, and deployment options including serverless hosting. Agentor can be used with various LLM providers and provides observability features for monitoring agent performance.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

• GetWeatherTool

  Get weather information for a specific location

  Try:What is the weather in London?Try:Check the weather forecast for Paris
• ShellTool

  Execute shell commands in a secure environment

  Try:Create a GIF using shell commandsTry:List files in the current directory
• get_weatherget_weather(location: str) -> str

  Get weather information for a specified location

  Try:Get the weather for New YorkTry:What's the weather like in Tokyo?

Comparable tools

Installation

pip install agentor

{
  "mcpServers": {
    "agentor": {
      "command": "python",
      "args": ["-m", "agentor.mcp"]
    }
  }
}

FAQ

What is the difference between LiteMCP and FastMCP?

LiteMCP is a native ASGI app that integrates directly with FastAPI using standard patterns, while FastMCP requires mounting as a sub-application, diverging from standard FastAPI primitives. LiteMCP also has built-in CORS support and allows for custom methods.

Can I use Agentor with different LLM providers?

Yes, Agentor supports various LLM providers including OpenAI (gpt-5-mini), Google (gemini/gemini-2.5-pro, gemini/gemini-3-flash-preview), and Anthropic (anthropic/claude-3.5).

On Hacker News

Recent discussion from the developer community.

• Show HN: Built an open-source SDK to simplify tool authentication for AI Agents▲ 1
  Story by theaniketmaurya · 2025-10-12

Compare agentor with