#3131st of 38 in Agent and MCP starters
mcp-agent
Python framework for building MCP-connected agents with composable workflow patterns
- Stars
- 8.6k
- License
- Apache-2.0
- Last commit
- Jan 2026
- Last release
- May 2025
- Language
- Python
8,074 on 31 Mar 20268,573 on 20 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
mcp-agent is a Python library that connects LLMs to MCP servers and manages server connection lifecycles. It implements the patterns from Anthropic's Building Effective Agents (parallel, orchestrator, router, evaluator-optimizer) as composable workflows. Execution runs on asyncio by default, or on Temporal for durable, resumable runs without API changes.
Who it is for: Python developers building MCP-based agents and workflows
Strengths
- Implements MCP tools, resources, prompts, OAuth, sampling, elicitation and roots
- Same decorators run on asyncio or Temporal for durable execution
- Agents can be exposed as MCP servers with a FastMCP-compatible API
- CLI scaffolds projects via uvx mcp-agent init
Weaknesses
- Last release was 2025-05-09, well behind recent commits
- Durable mode requires running a Temporal backend
- Managed cloud deploy (mcp-c) is labeled beta
- No Docker or compose files detected in the repo
What it needs
- Python, OpenAI, Anthropic, Google, Azure
- Needs Temporal (optional, durable execution), MCP servers (e.g. uvx, npx), LLM provider API key
Also in Agent and MCP starters
See all 38| Rank | Project | Score |
|---|---|---|
| 1 | deepseek-harnessPlugin-based agent harness from DeepSeek with a local Web UI | 70 out of 100 |
| 2 | mcp-typescript-templateTypeScript starter for remote MCP servers over HTTP with Express and Effect | 63 out of 100 |
| 3 | docker-agentDocker CLI plugin that runs multi-agent teams defined in YAML | 60 out of 100 |
| #4 | mcp-useTypeScript framework for building MCP servers with React Views and an Inspector | 59 out of 100 |
| #5 | computerDurable Object virtual filesystem with container, shell and JavaScript runtimes for agents | 59 out of 100 |