mcptool

Requirements

• Python 3.12+
• Node.js (for MCP tools)
• Playwright (installed via MCP tools)
• Environment variable: GROQ_API_KEY (for Groq LLM access)

Installation

1. Clone the repository:

   git clone <repo-url>
   cd mcptool

2. Install Python dependencies:

   pip install -r requirements.txt
   # or, if using poetry or pipenv, use your preferred tool

3. Install Node.js and Playwright (if not already installed):

   node --version
   npx playwright install chrome

Configuration

• MCP server configurations are stored in browser_mcp.json and similar files.
• Set your Groq API key in a .env file:

  GROQ_API_KEY=your_groq_api_key_here

Usage

Run the main interactive chat (with memory):

python main.py

• Type your queries to interact with the assistant.
• Type exit or quit to end the session.
• Type clear to reset the conversation history.

Run the simple MCP test script:

python main.py

• This will run a series of automated tests and then enter interactive mode.

Run the basic LLM test script:

python test.py

• This script tests basic LLM functionality and provides a simple chat interface.

Project Structure

• main.py - Main entry point for memory-enabled chat and MCP tool usage
• test.py - Basic LLM test and chat
• browser_mcp.json - MCP server configuration for Playwright and DuckDuckGo
• pyproject.toml - Python project metadata and dependencies
• .env - Environment variables (not committed)

Troubleshooting

• Ensure Node.js is installed and available in your PATH
• If Playwright is not found, run npx playwright install chrome
• Make sure your GROQ_API_KEY is set in the environment or .env file