EDUScan

🎯 Inspiration In many rural areas and underserved communities, students often struggle to understand complex academic content due to lack of access to quality teachers or tutoring resources. We wanted to build a simple, accessible solution that empowers any student to learn — directly from their own notes — using just a smartphone and the power of AI.

🤖 What it does EDUScan turns handwritten or printed school notes into simple explanations. Students just:

Upload or snap a photo of their notes.

EDUScan uses OCR to extract text.

Then, it calls the Perplexity Sonar API to explain the content — either normally or in Simplify Mode (for younger learners or easier comprehension).

In short: it reads, understands, and explains notes in seconds.

🛠️ How we built it Frontend: Built using Streamlit for fast and accessible UI.

OCR Engine: Used pytesseract to extract text from uploaded images.

LLM Integration: Connected to Perplexity's Sonar API for natural language explanations.

Environment Management: Used python-dotenv for API key management and secure handling.

Optional Mode Toggle: Implemented “Simplify Mode” to tailor explanations based on student level.

🧱 Challenges we ran into OCR Accuracy: Extracting clean text from handwritten notes with variable quality was tricky.

Prompt Engineering: Tuning prompts to get consistent and truly “simplified” responses from the LLM took experimentation.

API Limits: We had to work within rate limits of the Perplexity API, optimizing requests to avoid unnecessary calls.

🏆 Accomplishments that we're proud of Created a fully functional prototype in under 48 hours.

Made a tool that’s immediately useful for students with accessibility or learning challenges.

Demonstrated meaningful integration of computer vision + LLMs for real educational impact.

📚 What we learned The real-world challenges of OCR and language generation on varied inputs.

How to build for accessibility and inclusivity — especially for non-urban learners.

The importance of simplicity in UI/UX for educational tools.

Built With

• python
• sonar
• streamlit