LangGraph MCP Provider¶
This module provides integration with LangGraph, allowing conversion between MCP tools and LangGraph’s tool format.
Overview¶
The primary function in this module is to_langgraph_tools
, which takes a NostrMCPClient
instance and converts the available MCP tools into a list of tools compatible with LangGraph agents, such as those created with langgraph.prebuilt.create_react_agent
.
Usage¶
from langgraph.prebuilt import create_react_agent
from langchain_openai import ChatOpenAI
from agentstr import NostrMCPClient
from agentstr.mcp.providers.langgraph import to_langgraph_tools
# Assume nostr_mcp_client is an initialized and connected NostrMCPClient
async def setup_langgraph_agent(nostr_mcp_client: NostrMCPClient):
# Convert MCP tools to LangGraph format
langgraph_tools = await to_langgraph_tools(nostr_mcp_client)
# Create a LangGraph agent with the converted tools
agent = create_react_agent(
model=ChatOpenAI(),
tools=langgraph_tools,
prompt="You are a helpful assistant",
)
return agent
Note
For a complete, working example, check out the LangGraph Agent example.
Reference¶
See Also¶
LangGraph Provider — For using LangGraph agents with Agentstr.
agentstr.mcp.nostr_mcp_client.NostrMCPClient
— The client used to fetch MCP tools.