Study Bot – Generative AI Learning Assistant

An AI-powered Study Assistant API that answers academic questions in clear, simple language. Built with FastAPI, powered by Groq's LLaMA 3.1 8B (via LangChain), with persistent conversational memory backed by MongoDB Atlas.

🔗 Live API: https://limitbreak-ai.onrender.com 📘 Swagger UI: https://limitbreak-ai.onrender.com/docs

---

Overview

This project is a RESTful backend service that lets clients send a study-related question and receive an AI-generated, student-friendly explanation. It's deployed on Render as a production-style web service, with interactive API docs exposed via Swagger UI.

Features

• 🤖 Generative AI responses — powered by Groq's llama-3.1-8b-instant model
• 🧠 Conversational memory — retains context across multiple turns per user
• 💾 Persistent storage — chat history saved in MongoDB Atlas
• 📚 Study-focused — system prompt constrains the bot to academic/learning questions, answered in simple language
• ☁️ Cloud-deployed — live on Render with public API access
• 📖 Interactive docs — Swagger UI for testing without Postman/cURL

How Memory Works

Each interaction is stored in MongoDB with:

Field	Description
user_id	Unique identifier for the user/session
user_message	The user's query
bot_response	The AI-generated reply
timestamp	Order of the interaction

On each new request, prior messages for that user_id are retrieved and appended to the prompt sent to the LLM — simulating memory in an otherwise stateless API.

Tech Stack

• Framework: FastAPI
• LLM: Groq LLaMA 3.1 8B (via langchain-groq)
• Database: MongoDB Atlas (pymongo)
• Server: Uvicorn
• Deployment: Render
• Runtime: Python 3.12.3

API

POST /chat

Request body:

{
  "user_id": "student1",
  "message": "Explain Artificial Intelligence in simple words"
}

Response:

{
  "response": "Artificial Intelligence (AI) is like having a smart assistant..."
}

Try it live at /docs.

Local Setup

1. Clone the repo:

   git clone https://github.com/pavankarthikeyaatchyuta-lab/study-bot.git
   cd study-bot

2. Create a virtual environment and install dependencies:

   python -m venv venv
   source venv/bin/activate  # Windows: venv\Scripts\activate
   pip install -r requirements.txt

3. Create a .env file with:

   GROQ_API_KEY=your_groq_api_key
   MONGODB_URI=your_mongodb_connection_string
   

4. Run the server:

   uvicorn main:app --reload

5. Open http://127.0.0.1:8000/docs to test the /chat endpoint.

Deployment

Deployed on Render as a managed web service:

• Start command: uvicorn main:app --host 0.0.0.0 --port 10000 (see start.sh)
• Runtime: specified in runtime.txt
• Environment variables: GROQ_API_KEY, MONGODB_URI configured in Render dashboard

Future Improvements

• User-facing frontend interface
• Authentication and user management
• Rate limiting / input validation
• Analytics on chat usage
• Trimming/summarizing long chat histories before they hit the prompt