Serper MCP Client

Author: christseng89

README-Postman-Testing.md

@@ -37,3 +37,10 @@ docker build -t serper-mcp-server .
 docker run -it --rm -e SERPER_API_KEY=b3cf... serper-mcp-server
 docker run -it -d -e SERPER_API_KEY=b3cf... serper-mcp-server
 ```
+
+## Agent Client
+
+```cmd
+uv add langchain-openai
+uv add langgraph
+```

README-add.md

@@ -0,0 +1,86 @@
+import os
+import asyncio
+import traceback
+from dotenv import load_dotenv
+from langchain_openai import ChatOpenAI
+from langgraph.prebuilt import create_react_agent
+from langchain_mcp_adapters.client import MultiServerMCPClient
+from langchain_core.messages import AIMessage, BaseMessage, HumanMessage
+
+load_dotenv()
+
+SERPER_API_KEY = os.getenv("SERPER_API_KEY")
+
+
+class SerperAgent:
+    def __init__(self) -> None:
+        self.llm = ChatOpenAI(model="gpt-4o-mini", temperature=0)
+        self.client: MultiServerMCPClient | None = None
+        self.agent = None
+        self.is_initialized = False
+
+    async def initialize(self) -> None:
+        """Initialize MCP client and agent"""
+        try:
+            # Initialize MCP client for serper-mcp-server
+            self.client = MultiServerMCPClient(
+                {
+                    "serper": {
+                        "transport": "stdio",
+                        "command": "uvx",
+                        "args": ["serper-mcp-server"],
+                        "env": {"SERPER_API_KEY": SERPER_API_KEY},
+                    }
+                }
+            )
+            tools = await self.client.get_tools()
+            self.agent = create_react_agent(self.llm, tools)
+            self.is_initialized = True
+            print("✅ MCP client initialized with serper-mcp-server.")
+        except Exception:
+            traceback.print_exc()
+            raise
+
+    async def ask(self, messages: list[BaseMessage]) -> str:
+        """Ask a question to the agent"""
+        try:
+            if not self.is_initialized:
+                await self.initialize()
+
+            result = await self.agent.ainvoke({"messages": messages})
+            for msg in reversed(result["messages"]):
+                if isinstance(msg, AIMessage):
+                    return msg.content
+            return "No valid AI response."
+        except Exception as e:
+            traceback.print_exc()
+            return f"Error: {e}"
+
+    async def close(self):
+        """Clean up"""
+        # MultiServerMCPClient does not have close(), just clear the client
+        self.client = None
+        self.agent = None
+        self.is_initialized = False
+        print("🛑 MCP client closed.")
+
+
+async def main():
+    agent = SerperAgent()
+
+    # Example questions
+    questions = [
+        "What is the weather in Taipei?",
+        "What is the benefit of using FastMCP?"
+    ]
+
+    for q in questions:
+        messages = [HumanMessage(content=q)]
+        response = await agent.ask(messages)
+        print(f"💡 Q: {q}\n➡️ A: {response}\n")
+
+    await agent.close()
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

fastmcp_serper_client.py

@@ -11,7 +11,14 @@ requires-python = ">=3.11"
 dependencies = [
     "aiohttp>=3.11.16",
     "certifi>=2025.1.31",
+    "fastmcp>=2.10.6",
+    "langchain>=0.3.27",
+    "langchain-mcp-adapters>=0.1.9",
+    "langchain-openai>=0.3.28",
+    "langgraph>=0.6.2",
+    "langgraph-prebuilt>=0.6.2",
     "mcp[cli]>=1.6.0",
+    "openai>=1.98.0",
     "pydantic>=2.11.2",
     "python-dotenv>=1.1.0",
 ]