# lastmile-ai/mcp-agent

Build effective agents using Model Context Protocol and simple workflow patterns

Repository: https://github.com/lastmile-ai/mcp-agent
Canonical: https://ross.abutalabs.com/products/mcp-agent
Language: Python
License: Apache-2.0
License Family: permissive
Topics: agents, ai, ai-agents, llms, mcp, model-context-protocol, python, llm
Last push: 2026-01-25T16:35:16+00:00

## Health v2 (maintenance only)
Score: 52/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 64, release rhythm 40, longevity 44
- inputs: {"age_days": 624, "days_push": 220, "days_rel": 481, "gap_med": 8.0, "n_releases_24m": 9}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 8519, forks 880 (observed 2026-08-28T04:10:23.307631+00:00)

## What it is
mcp-agent is a Python framework for building AI agents on top of the Model Context Protocol (MCP) using simple, composable workflow patterns. It handles MCP server connection lifecycles, implements the patterns from Anthropic's 'Building Effective Agents' essay, and supports durable, pausable/resumable agent workflows backed by Temporal.

## Use cases
- build ai agents using the model context protocol in python
- implement anthropic's building effective agents patterns like routing and parallelization
- manage mcp server connections and lifecycles for an llm agent
- create durable resumable agent workflows with temporal
- orchestrate multiple mcp servers in a single agent application
- chain llm calls into composable multi-step agent workflows
- build a multi-agent system with evaluator and orchestrator patterns

## When to choose
- you are building agents in Python that connect to MCP servers and want connection lifecycle handled for you
- you want ready-made, composable implementations of Anthropic's effective agent patterns rather than hand-rolling orchestration
- you need agents that can pause, resume, and recover via durable execution on Temporal without changing your agent API
- you want a lightweight SDK that scales from simple single-agent scripts to sophisticated workflows

## When to avoid
- your project is not in Python, since mcp-agent is a Python-only SDK
- you just need a single LLM API call with no MCP tool integration or workflow orchestration
- you want a turnkey chatbot product with UI rather than a framework for writing agent code
- you cannot run or access a Temporal service and require durable workflow features

## Facets
- artifact type: framework
- maturity: active
- function: agent-framework, mcp, workflow-automation, llm-inference
- domain: artificial-intelligence, large-language-models, developer-tools
- platform: python, cross-platform
- tags: model-context-protocol, temporal, durable-execution, agent-patterns, llm-orchestration, anthropic, openai, ai-agents

## Member repositories
- lastmile-ai/mcp-agent (main) score 52

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:23.307631+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:26:40.753154+00:00, confidence not recorded.
  - readme: https://github.com/lastmile-ai/mcp-agent (fetched 2026-08-28T04:10:23.307631+00:00, sha 4f5722d21f3b)
- Data as of 2026-08-30T08:39:29.467469+00:00.