LangChain and LangGraph

The langchain-recruitmcp package connects to the RecruitMCP MCP server and returns its tools as LangChain tools. It reads your key from RECRUITMCP_KEY. Get a key in the console.

Install

pip install "langchain>=1.0" langchain-openai langchain-recruitmcp

langchain is not a dependency of the toolkit itself. Install it for the agent example: create_agent is in langchain.agents from langchain 1.0.

List the tools

import asyncio
from langchain_recruitmcp import RecruitMCPToolkit

async def main():
    tools = await RecruitMCPToolkit().aget_tools()
    print([tool.name for tool in tools])

asyncio.run(main())

Run an agent

import asyncio
from langchain.agents import create_agent
from langchain_recruitmcp import RecruitMCPToolkit

async def main():
    tools = await RecruitMCPToolkit().aget_tools()
    agent = create_agent("openai:gpt-4.1-mini", tools=tools)
    result = await agent.ainvoke({"messages": [{"role": "user", "content": "Find backend engineers in Ghent, Belgium. Dry run only."}]})
    print(result["messages"][-1].content)

asyncio.run(main())

This needs OPENAI_API_KEY. Any LangChain chat model works. Every await sits inside main(), which asyncio.run starts, so the file runs as a plain script.

Options

Argument	Default	What it does
api_key	RECRUITMCP_KEY	Your RecruitMCP key
sourcing_only	True	Leaves out outreach tools
include_billing	False	Leaves out billing tools such as set_auto_refill unless True (version 0.1.1 and later)
include	None	Keep only these tool names
exclude	None	Drop these tool names

On version 0.1.0, pass exclude=["set_auto_refill", "get_auto_refill"] to keep billing tools away from the agent. The tools also depend on the key's scopes: the server lists only the tools the key allows.