MCP Client logo

MCP Client

by rakesh-eltropy

A simple REST API and CLI client to interact with Model Context Protocol (MCP) servers. It leverages LangChain to execute LLM prompts and enables multiple MCP servers to collaborate on queries.

View on GitHub

Last updated: N/A

What is MCP Client?

The MCP Client is a tool that allows you to interact with Model Context Protocol (MCP) servers through a REST API or a command-line interface. It facilitates communication with various LLMs and MCP-compatible databases, allowing for complex queries and data retrieval.

How to use MCP Client?

To use the client, clone the repository, navigate to the project directory, set the necessary API keys (OPENAI_API_KEY, BRAVE_API_KEY) either as environment variables or in the mcp-server-config.json file. Then, either run the CLI using uv run cli.py or start the REST API server with uvicorn app:app --reload. The CLI provides a chat command for interacting with the LLM, and the REST API can be accessed via HTTP requests.

Key features of MCP Client

  • MCP-Compatible Servers support

  • Integrated with LangChain

  • LLM Provider Support for OpenAI, Claude, Gemini, AWS Nova, Groq, Ollama, etc.

  • REST API and CLI Client

Use cases of MCP Client

  • Querying data from multiple MCP servers simultaneously

  • Integrating LLMs with local databases for knowledge retrieval

  • Building applications that require function-based LLM interactions

  • Automating data retrieval and analysis tasks with LLMs

FAQ from MCP Client

What is MCP?

MCP stands for Model Context Protocol. It is a protocol for enabling interactions with MCP-compatible servers.

Which LLM providers are supported?

The client supports any LLM provider that offers a function-based API, including OpenAI, Claude, Gemini, AWS Nova, Groq and Ollama.

How do I configure the API keys?

API Keys can be configured directly in the mcp-server-config.json file, or can be set as environment variables.

What is Langchain?

Langchain is a framework that allows easy set up of LLM prompts.

How do I add my own MCP server?

Additional MCP servers can be added in the mcp-server-config.json file.