What is Mcp Client

MCP-Client is a software tool that allows users to connect to a local MCP server and interact with it using two different integrations: OpenAI and Claude. It utilizes standard input/output for communication.

Use cases

Use cases for MCP-Client include developing AI-powered applications, conducting experiments with different language models, and integrating AI functionalities into existing software solutions.

How to use

To use MCP-Client, clone the repository, install the dependencies using npm, compile the TypeScript files, set up the .env file with the appropriate API keys, and run the client for either OpenAI or Claude by specifying the path to the MCP server script.

Key features

Key features of MCP-Client include support for both OpenAI and Claude integrations, automatic detection of server script types (Python or JavaScript), and interactive streaming of tool results and LLM responses.

Where to use

MCP-Client can be used in various fields such as software development, AI research, and any application that requires interaction with language models hosted on an MCP server.

MCP Client (OpenAI and Claude)

This repo contains two separate MCP clients — one is MCP Client with OpenAI Integration and another one is with Claude Integration — that can connect to local MCP server and handle interactions. The client uses standard input/output (stdio) to talk to the MCP server.

Directory Structure

.
├── .git/
├── src/
│   ├── mcp-client-openai.ts
│   └── mcp-client-claude.ts
├── package.json
├── tsconfig.json
└── README.md

Setup Instructions

1. Clone the Repository

git clone https://github.com/devgotomarket/MCP-Client.git
cd MCP-Client

2. Install dependencies

npm install

3. Compile TypeScript

npm run build

This will generate the compiled JS files under the build/ directory.

4. prepare .env

Make sure your .env file is set up with the appropriate API key depending on the client:

# For OpenAI
OPENAI_API_KEY=your_openai_key_here

OPENAI_MODEL=openai_model ex:(gpt-4o)

# For Claude
ANTHROPIC_API_KEY=your_anthropic_key_here

ANTHROPIC_MODEL=anthropic_model ex:(claude-3-5-sonnet-20241022)

5. Run MCP Client (LLM-OpenAI)

node build/mcp-client-openai.js path/to/server.py

6. Run Claude Client (LLM-Claude)

node build/mcp-client-claude.js path/to/server.py

Replace path/to/server.py with the path to the MCP server script you want to connect to — it can be local or remote.

Notes

• You can use either .py (Python) or .js (Node) server scripts.
• The client will automatically detect the type and establish the transport.
• Tool results and LLM responses are streamed in an interactive prompt.