mcp-client

MCP.Pizza Chef: vurtnec

The MCP Client is a Node.js-based orchestrator that manages connections to MCP servers and tools, enabling advanced AI model interactions. It registers servers, lists available tools, and facilitates tool invocation through a RESTful API. Designed for developers, it simplifies integrating and managing multiple MCP servers and tools in AI-enhanced workflows, supporting real-time context flow and multi-step reasoning across providers.

Use This MCP client To

Register and manage multiple MCP servers automatically List available tools for a specific MCP server Invoke MCP tools via RESTful API calls Configure and customize MCP client settings for different environments Integrate MCP client into AI workflows for real-time context management

README

MCP Client

Overview

This is a Model Context Protocol (MCP) client project designed to facilitate advanced interactions with AI models and services.

Prerequisites

  • Node.js (version 16 or higher)
  • pnpm (package manager)

Installation

  1. Clone the repository
git clone https://github.com/your-username/mcp-client.git
cd mcp-client
  1. Install dependencies
pnpm install

Configuration

  1. Copy mcp_config.json.example to mcp_config.json
  2. Modify the configuration file with your specific server and tool settings

Usage

  1. To run the client:
pnpm start
  1. All your own servers will be registered when you run the client.

  2. You can use the list-tools API to list all the tools available for a specific server.

curl --location 'http://localhost:3000/list-tools/{serverName}'
  1. You can use the call-tool API to call a specific tool.
curl --location 'http://localhost:3000/call-tool' \
--header 'Content-Type: application/json' \
--data '{
    "serverName": "server_name",
    "toolName": "tool_name",
    "args": {
        // Arguments must match the inputSchema from the list-tools response
        // Required fields must be included
        // Optional fields can be omitted
        // Example based on above schema:
        // "issue_key": "PROJ-123"  // Required
        // "expand": "..."          // Optional
    }
}'

Cursorrules

All tool calls will be logged in the .cursorrules file.

IMPORTANT:

  • YOU MUST COPY CURSORRULES TO YOUR OWN REPO AND EDIT IT TO YOUR OWN NEEDS.
  • YOU MUST COPY "mcp_config.json" TO YOUR OWN REPO AND EDIT IT TO YOUR OWN NEEDS.
  • YOU MUST USE COMPOSER WITH AGENT.

Contributing

  1. Fork the repository
  2. Create your feature branch (git checkout -b feature/AmazingFeature)
  3. Commit your changes (git commit -m 'Add some AmazingFeature')
  4. Push to the branch (git push origin feature/AmazingFeature)
  5. Open a Pull Request

License

Distributed under the MIT License. See LICENSE for more information.

Contact

Your Name - your.email@example.com

Project Link: https://github.com/your-username/mcp-client

mcp-client FAQ

How do I install the MCP client?
Clone the repository, then run 'pnpm install' with Node.js 16+ and pnpm installed.
How do I configure the MCP client?
Copy 'mcp_config.json.example' to 'mcp_config.json' and modify server and tool settings as needed.
How do I start the MCP client?
Run 'pnpm start' in the project directory to launch the client and register servers.
How can I list tools available on a server?
Use the 'list-tools' API endpoint with the server name to retrieve available tools.
How do I call a specific tool?
Use the 'call-tool' API endpoint with the appropriate tool parameters in a POST request.
What prerequisites are needed?
Node.js version 16 or higher and pnpm package manager are required.
Can the MCP client handle multiple servers?
Yes, it automatically registers and manages multiple MCP servers.
Is the MCP client limited to a specific AI model provider?
No, it is provider-agnostic and works with models like OpenAI, Claude, and Gemini.