Agent-MCP logo

Agent-MCP

by Zeppelinpp

Agent-MCP is a local Model Context Protocol (MCP) server-client agent. It implements a local web search MCP server using Google SerpAPI and an MCP client that can call both the local server and an external MCP server provided by Simthery.

View on GitHub

Last updated: N/A

What is Agent-MCP?

Agent-MCP is a local implementation of the Model Context Protocol (MCP) for web search. It consists of a server that uses Google SerpAPI to perform searches and a client that can interact with both the local server and external MCP servers.

How to use Agent-MCP?

To use Agent-MCP, first deploy the MCP server locally. Then, use the provided MCP client to interact with the server, specifying the desired search query. The client can be configured to use OpenAI's API for interaction and includes JSON validation.

Key features of Agent-MCP

  • Local MCP Server

  • Google SerpAPI integration

  • MCP Client

  • OpenAI API compatibility

  • Simthery platform support

  • JSON validation

Use cases of Agent-MCP

  • Web search integration for LLMs

  • Local testing of MCP implementations

  • Agent development with web search capabilities

  • Retrieval-augmented generation (RAG)

  • News aggregation

FAQ from Agent-MCP

What is MCP?

MCP stands for Model Context Protocol. It's a protocol for agents to interact with tools and services.

What is Google SerpAPI?

Google SerpAPI is an API that allows you to extract search engine results from Google.

What is Simthery?

Simthery is a platform that provides an external MCP server service.

Which LLMs are compatible?

The client is tested and works well with Qwen2.5 series models and DeepSeek-V3.

How can I run the client?

You can run the client using the provided python script: python ./mcp-client/openai-agent.py