Connect MCP tools

Model Context Protocol (MCP) is an open standard for connecting AI agents to external services. Instead of writing custom integration code, you point the agent at an MCP server and it discovers all available tools automatically.

Popular MCP servers: GitHub, Notion, Slack, Postgres, filesystem, web search, and hundreds more — all work with the same two lines of code.

Connect a single server

from gantrygraph import GantryEngine
from gantrygraph.mcp import MCPClient
from langchain_anthropic import ChatAnthropic

agent = GantryEngine(
    llm=ChatAnthropic(model="claude-sonnet-4-6"),
    tools=[
        MCPClient("npx -y @modelcontextprotocol/server-filesystem /workspace"),
    ],
    max_steps=20,
)

result = agent.run("List every file modified today and show the 3 largest.")
print(result)

MCPClient starts the server subprocess when the agent run begins and terminates it when it ends — even on errors.

Connect multiple servers

Use MCPToolRegistry to give the agent tools from several servers at once. All tools are flattened into one list and de-duplicated by name.

from gantrygraph.mcp import MCPClient, MCPToolRegistry

registry = MCPToolRegistry([
    MCPClient("npx -y @modelcontextprotocol/server-filesystem /workspace"),
    MCPClient("npx -y @modelcontextprotocol/server-github"),
    MCPClient("npx -y @modelcontextprotocol/server-postgres postgresql://..."),
])

agent = GantryEngine(
    llm=ChatAnthropic(model="claude-sonnet-4-6"),
    tools=[registry],
    max_steps=30,
)

result = agent.run(
    "Find all open PRs tagged 'bug' in the repo, "
    "check which ones have failing tests in the database, "
    "and write a summary to /workspace/report.md."
)

Mix MCP tools with custom tools

You can combine MCPClient with @gantry_tool functions and built-in action groups:

from gantrygraph import GantryEngine, gantry_tool
from gantrygraph.mcp import MCPClient
from gantrygraph.actions import FileSystemTools

@gantry_tool
async def notify_slack(message: str) -> str:
    """Post an urgent notification to #alerts on Slack."""
    await slack.post("#alerts", message)
    return "posted"

agent = GantryEngine(
    llm=...,
    tools=[
        MCPClient("npx -y @modelcontextprotocol/server-github"),
        FileSystemTools(workspace="/workspace"),
        notify_slack,
    ],
    max_steps=20,
)

Use MCP tools directly (without an agent)

If you just want to list or call tools from an MCP server programmatically:

import asyncio
from gantrygraph.mcp import MCPClient

async def main():
    async with MCPClient("npx -y @modelcontextprotocol/server-filesystem /tmp") as client:
        tools = client.get_tools()
        print([t.name for t in tools])
        # ['read_file', 'write_file', 'list_directory', ...]

asyncio.run(main())

Finding MCP servers

Browse the official registry at modelcontextprotocol.io/servers or search npm for packages starting with @modelcontextprotocol/.

Any MCP-compatible server works — including ones you build yourself.