Day 4 of my AI Engineering Journey β Built a personal AI that answers questions about YOUR own study notes.
When you study, you take notes. But finding information across many notes is slow. And testing yourself on what you wrote is even harder.
This app solves both problems:
- Save your notes in one place
- Ask AI anything about YOUR notes
- AI answers ONLY from what YOU wrote
Most AI answers from its general training data. This AI answers ONLY from your personal notes.
That is the core idea behind RAG β Retrieval Augmented Generation.
This project is the simplest version of RAG possible: Your notes β injected into system prompt β AI answers from them.
- π Add notes β save any study topic with full content
- π View all notes β see everything you have saved
- π€ Ask AI about all notes β get answers across all topics
- π Ask AI about one note β deep dive into a specific topic
- π Search notes β find notes by keyword instantly
- ποΈ Delete notes β remove notes you no longer need
- πΎ Persistent storage β notes saved to JSON, survive app restarts
system_prompt = f"""You are a study assistant.
Answer ONLY based on these notes:
{user_notes} # β This is the magic
"""By injecting the user's notes directly into the system prompt, the AI becomes an expert on exactly what YOU studied. Nothing more. Nothing less.
Notes are saved locally in notes.json:
[
{
"id": 1,
"topic": "RAG Systems",
"content": "RAG stands for Retrieval Augmented Generation...",
"date": "2026-05-11 14:30"
}
]| Tool | Purpose |
|---|---|
| Python | Core programming language |
| Groq API | AI inference platform |
| LLaMA 3.3 70B | The underlying language model |
| JSON | Local data storage |
| python-dotenv | Secure API key management |
| Git & GitHub | Version control |
- Clone the repository
git clone https://github.com/PrashikSawant/ai-study-notes.git
cd ai-study-notes- Install dependencies
pip install groq python-dotenv- Create a
.envfile
GROQ_API_KEY = "Add your API KEY here"- Run the app
python main.py- How to read and write JSON files in Python
- How context injection works β the foundation of RAG
- How to build a multi-feature menu driven application
- Why AI answers change completely based on what context you give it
- How persistent local storage works without a database
- π Research Paper Summarizer
- π PDF Reader + Q&A with ChromaDB
- π Full RAG System with Vector Database
I am Prashik β an aspiring AI Engineer on a 4-month intensive journey to become job-ready in Generative AI Engineering.
Follow my journey β LinkedIn
See all projects β GitHub
Day 4 of 120 | May 11, 2026 | Status: β Complete