mcp-agent

mcp-agent

Build effective agents using Model Context Protocol and simple workflow patterns.

MCPDevOpen sourceMCP代理工作流
Type
MCP
Transport
stdio
Open source
Yes
GitHub Stars
★ 8.4k
Source
mcp-github

Overview

mcp-agent is a simple, composable framework for building effective agents using the Model Context Protocol (MCP). It fully supports MCP and handles the tedious work of managing MCP server connection lifecycles. mcp-agent implements all the patterns described in Anthropic's 'Building Effective Agents' guide, implemented in a composable way that allows you to chain these patterns together. Additionally, it supports persistent execution, suitable for everything from simple agents to complex workflows based on Temporal. Ideal for developers looking to build efficient and reliable agent applications.

Capabilities

  • Full MCP protocol support
  • Implementation of multiple effective agent patterns
  • Support for persistent execution
  • Scalable to production-grade workloads

Use cases

Building complex multi-step agentsIntegrating LLMs with external data sourcesCreating and deploying chatbot applicationsHandling large-scale concurrent requests

Setup

Requires: API KeyNode 环境
pip install mcp-agent

This information was compiled by AI from public sources and may contain inaccuracies — please refer to the source.

FAQ

How do I get started with mcp-agent?

See the quick start guide: https://docs.mcp-agent.com/get-started/overview

What MCP features does mcp-agent support?

Supports core features such as tools, resources, prompts, notifications, and advanced features like OAuth, sampling, and more.

Related skills