LearnAble

Inspiration

I was inspired to create LearnAble when I learned that students with dyslexia or visual processing disorders just don't seem to be able to access regular classroom materials. There are many educational tools that focus on digitization without filling the need for real accessibility. I wanted to create something that could fill that gap—a device that not only simplifies the learning process but also adapts to the varied needs of learners using ethical AI.

What it does

LearnAble is an artificial intelligence-powered web application that allows students to input class notes, lectures, or scanned documents and turn them into simpler, accessible study materials. The application does raw text extraction through optical character recognition (OCR), summarizes it to plain English bullet points through natural language processing (NLP), and generates voice narration to support the auditory learner. It also uses machine learning to detect and display major topics, making it easier for users to learn valuable concepts. Every user can securely sign up for their own account, log in, and bookmark personal notes to come back to at any time, which makes it a continuous learning process.

How I built it

I created LearnAble using Google Colab and Gradio as the primary development environments. The backend is composed of a summary pipeline using Hugging Face Transformers' T5 model, OCR using PyTesseract, and speech synthesis with gTTS. Keyword frequency is analyzed by scikit-learn's CountVectorizer, plotted using Matplotlib. The UI is implemented through Gradio's tab system using a soft blue and white theme, with animation and custom styles for accessibility. A JSON-based authentication system allows users to securely sign up, log in, and save their progress.

Challenges I ran into

One of the biggest challenges I faced was integrating multiple AI models and workflows into a single, light UI without sacrificing functionality or accessibility. Keeping accuracy and context-relevance of the text summarization intact was another challenging task. Creating a login system that fit well within a Colab-based environment while keeping user data intact was also something I had to achieve.

Accomplishments that I'm proud of

Despite the challenges, I'm pleased with the final product. I managed to build a full-stack, AI-powered accessibility tool inside a Google Colab notebook from the ground up. LearnAble is not just a technical project—it's a huge step towards accessible education for all. I'm especially proud of the way the app provides users with a smooth experience and yet packs in advanced machine learning features.

What I learned

Through this project, I gained experience in building end-to-end AI applications, integrating multiple open-source libraries seamlessly, and designing for actual users with specific accessibility needs. I gained a better sense of ethical design in AI and how to use these technologies responsibly.

What's next for LearnAble

In the future, I would like to expand LearnAble further with multilingual translation support for non-native English speaking students. I would also like to implement collaboration features, such as annotation and group study functionality, and ultimately build a more mature backend with database integration to enable user progress and note history storage. As it continues to be perfected, I believe LearnAble has the potential to be a one-stop-shop for inclusive and adaptive learning.

Built With

• googlecolab
• gradio
• gtts
• json
• matplotlib
• pytesseract
• python
• scikit-learn
• t5