MCP-PuchAI-Hackathon by GVKarthik-dev - MCP Server

`#BuildWithPuch`

🧠 Groq AI Knowledge, Document Q&A & Health Assistant MCP

This is a Modular Conversational Platform (MCP) powered by FastMCP, Groq LLaMA 3.1 (8B) via ChatGroq, and integrated with document parsing, health tools, and AI assistants for various domains including:

✅ Health checks
📄 Document Q&A (PDF/DOCX)
📘 Knowledge Assistant
🥗 Diet & Nutrition
🧠 Mental Health Support
🚨 Emergency Instructions
🏋️‍♂️ Exercise & Fitness Guidance

🚀 Features

✅ AI-generated health responses with safety disclaimers.
📄 Upload and query documents (PDF, DOCX) in natural language.
🔍 Ask general knowledge questions.
🧘 Mental health and emotional support.
🥗 Nutrition tips and diet planning.
🆘 First-aid emergency response instructions.
🏃‍♂️ Exercise and fitness recommendations.
🧩 Modular design with clean prompt builders and async tools.

🛠️ Setup Instructions

1. Clone the repository

git clone https://github.com/GVKarthik-dev/MCP-PuchAI-Hackathon.git
cd MCP-PuchAI-Hackathon

2. Install dependencies

Note: Make sure Python 3.9+ is installed & Install uv

uv sync

3. Create Environment File

Create a .env file in the root directory with your Groq API key and any other required secrets:

GROQ_API_KEY=YOUR_GROQ_API_KEY

4. Run the server

uv run python Server.py

The server will start at: http://localhost:8000

🧰 Tools Overview

Tool	Description
`health_check`	Answers user health queries.
`upload_and_qa`	Parses uploaded documents and answers questions contextually.
`ask_knowledge`	Provides detailed explanations of any topic.
`diet_and_nutrition`	Offers advice and tips on nutrition and healthy eating.
`mental_health_support`	Responds with emotional support and coping strategies.
`emergency_instructions`	Gives step-by-step first-aid guidance for emergencies.
`exercise_and_fitness`	Provides fitness recommendations and recovery tips.

🧱 Tech Stack

FastMCP - Modular conversational framework
ChatGroq - Integration with Groq's blazing fast LLaMA 3.1 models
pdfplumber - Extracts text from PDFs
python-docx - Reads .docx documents
dotenv - Loads environment variables
eunomia_mcp - Eunomia policy-based middleware

🧪 Example Usage

Health Question

{
  "tool": "health_check",
  "query": "What should I do if I have a mild fever?"
}

Document Q&A

{
  "tool": "upload_and_qa",
  "doc_base64": "<base64-encoded-pdf-or-docx>",
  "question": "What are the main responsibilities of the company mentioned?",
  "file_type": "pdf"
}

License

This project is licensed under the MIT License - see the file for details.

⚠️ Disclaimers

The health and mental support information is not a replacement for professional medical advice.
Emergency instructions are intended for first-aid and awareness, not as certified training.

🧩 Customization

You can add more tools or prompts by extending:

Prompt builders (functions like build_*_prompt)
Tool functions (@mcp.tool)

All prompt generation is modular and easy to modify.

👨‍💻 Author

Developed with ❤️ using FastMCP and Groq LLMs.