Inspiration
As students, we often struggle with incomplete notes—missed classes, fast lectures, unclear screenshots, or last-minute exam preparation. This inspired us to create SmartFill Notes, an AI-powered mobile app that can automatically fill missing text in incomplete notes directly on an Arm-based device. We wanted a tool that saves time, improves learning, and uses AI in a meaningful way.
🎯 What It Does
SmartFill Notes detects missing words or lines in class notes and intelligently fills the gaps using contextual on-device AI. It helps students:
Complete incomplete notes instantly
Understand content better with AI-generated explanations
Scan handwritten text and auto-complete it
Export clean notes as PDF
Learn faster with a simple and friendly interface
🛠️ How We Built It
We built SmartFill Notes as an Android application running on Arm architecture. Our development steps included:
Designing the UI in Android Studio
Training lightweight AI models using TensorFlow Lite
Optimizing the model to run efficiently on Arm-based mobile CPUs
Integrating OCR for extracting text from handwritten notes
Implementing a gap-filling algorithm using context analysis
Running multiple on-device tests to ensure stability and accuracy
We ensured that all processing runs locally on the device for speed, privacy, and alignment with the Hackathon requirements.
📚 What We Learned
Throughout this project, we learned:
How to optimize AI models for Arm processors
Building real on-device AI apps using TFLite
Efficient text processing and gap detection
UI/UX design for educational apps
Managing GitHub repositories with open-source licensing
Creating demo-ready mobile applications
🚧 Challenges We Faced
Some of the main challenges included:
Making the AI model lightweight enough to run smoothly on mobile
Achieving accurate gap prediction from messy or incomplete notes
Converting handwritten text into usable input
Ensuring the AI-generated sentences remain meaningful and educational
Fine-tuning performance without external servers or cloud computation
Each challenge pushed us to improve our design, research better techniques, and create an app that is fast, reliable, and accessible.
🚀 What's Next
We aim to expand SmartFill Notes with:
Diagram label completion
Multilingual note support
Offline dictionary-based enhancement
Collaborative note correction
Personalized note difficulty suggestions
Log in or sign up for Devpost to join the conversation.