Use with LangChain

LangChain is a popular framework for building applications with large language models (LLMs).

1. Install Required Dependencies:

   pip install langchain langchain-anthropic

2. Configure LangChain with MCP:

   from langchain_anthropic import ChatAnthropic
   from langchain.tools import tool
   from langchain.agents import AgentExecutor, create_structured_chat_agent
   from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
    
   # Initialize the Claude model
   model = ChatAnthropic(model="claude-3-sonnet-20240229")
    
   # Define a function to connect to your MCP server
   @tool
   def connect_to_mcp_server(query: str) -> str:
       """
       Connects to the MCP server at https://api.sapience.xyz/mcp and processes the query.
       Args:
           query: The query to process through the MCP server
       Returns:
           The response from the MCP server
       """
       # Implement connection to your MCP server
       # This is a simplified example - you'll need to implement the actual HTTP request
       import requests
       response = requests.post(
           "https://api.sapience.xyz/mcp",
           json={"query": query}
       )
       return response.json()
    
   # Create a list of tools
   tools = [connect_to_mcp_server]
    
   # Define the prompt for the agent
   prompt = ChatPromptTemplate.from_messages([
       ("system", "You are a helpful assistant."),
       ("human", "{input}"),
       MessagesPlaceholder(variable_name="agent_scratchpad")
   ])
    
   # Create an agent with the tools and prompt
   agent = create_structured_chat_agent(model, tools, prompt)
    
   # Create an agent executor
   agent_executor = AgentExecutor(agent=agent, tools=tools, verbose=True)
    
   # Execute the agent
   result = agent_executor.invoke({"input": "Process this data using the MCP server tools"})
   print(result["output"])

3. Advanced Integration: For more advanced integration, you can create custom Tool classes that specifically target different functionalities provided by your MCP server.

   from langchain.pydantic_v1 import BaseModel, Field
    
   class MCPToolInput(BaseModel):
       parameter1: str = Field(..., description="Description of parameter1")
       parameter2: int = Field(..., description="Description of parameter2")
    
   @tool(args_schema=MCPToolInput)
   def mcp_specific_tool(parameter1: str, parameter2: int) -> str:
       """
       A specific tool that uses the MCP server for a particular function.
       """
       # Implement specific functionality
       pass

For more information on LangChain's tool and agent capabilities, refer to the LangChain documentation.