MCP to LangChain/LangGraph Adapter
About
Addapter that turns MCP server tools into langchain usable tools
Explore
- Connects to any MCP server via stdio transport.
- Automatically discovers and converts MCP tools to LangChain BaseTool.
- Supports both LangChain agents and LangGraph react agents.
- Manages conversation history with LangGraph memory checkpoints.
- Allows custom environment variables per server process.
To use this adapter, you need to have the necessary packages installed:
load_dotenv()
Alternatively, you can set the API key directly in your environment or code:
```python
import os
os.environ["OPENAI_API_KEY"] = "your_api_key_here"
config = {"configurable": {"thread_id": "example-thread"}}
Here's a complete example of how to use the adapter:
from mcp_langchain_adapter import MCPAdapter
name
Name of the tool
description
Description of the tool
server_script_path
Path to the MCP server script
env
Optional environment variables for the server process
args_schema
Optional Pydantic model for tool arguments
print(f"Found {len(tools)} tools:")
for tool in tools:
print(f"- {tool.name}: {tool.description}")
Once you have the tools, you can use them in LangChain applications:
from langchain.agents import AgentExecutor, create_react_agent
from langchain.prompts import PromptTemplate
from langchain_openai import ChatOpenAI
agent = create_react_agent(llm, tools, prompt_template)
agent_executor = AgentExecutor(agent=agent, tools=tools, verbose=True)
LangGraph provides a more modern, flexible approach to building agents. Here's how to use our MCP tools with LangGraph:
``python
from langchain_openai import ChatOpenAI
from langchain_core.messages import HumanMessage
from langgraph.prebuilt import create_react_agent
from langgraph.checkpoint.memory import MemorySaver
agent = create_react_agent(
llm,
tools,
prompt="You are a helpful AI assistant that can use tools to solve problems.",
checkpointer=memory
)
The MCPToolWrapper class extends LangChain's BaseTool to wrap MCP tools:`
python`
MCPToolWrapper(
name: str,
description: str,
server_script_path: str,
env: Optional[Dict[str, str]] = None,
args_schema: Optional[Type[BaseModel]] = None
)
- : Name of the tooldescription
- : Description of the toolserver_script_path
- : Path to the MCP server scriptenv
- : Optional environment variables for the server processargs_schema
- : Optional Pydantic model for tool arguments
print(f"Found {len(tools)} tools:")
for tool in tools:
print(f"- {tool.name}: {tool.description}")
add_tool = adapter.get_tool_by_name("add")
if add_tool:
result = add_tool.run({"a": 5, "b": 7})
print(f"Result of add(5, 7): {result}")
weather_tool = adapter.get_tool_by_name("get_weather")
if weather_tool:
result = weather_tool.run({"city": "London"})
print(f"Result of get_weather('London'): {result}")
``
Table of Contents
- Introduction - Installation - Getting Started - Setting Up the MCP Server - Connecting to the MCP Server - Using MCP Tools with LangChain - Using MCP Tools with LangGraph - API Reference - MCPAdapter - MCPToolWrapper - Utility Functions - Examples - Basic Usage - Integration with LangChain Agents - Integration with LangGraph Agents - Troubleshooting - ContributingIntroduction
The MCP to LangChain/LangGraph Adapter bridges the gap between MCP servers, which provide various tools through a standardized interface, and LangChain/LangGraph, popular frameworks for building applications with large language models. This adapter enables you to: - Connect to an MCP server - Discover available tools - Convert MCP tools to LangChain-compatible tools - Use these tools in LangChain agents, chains, and LangGraph agentsInstallation
To use this adapter, you need to have the necessary packages installed: ```bashSign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
