Cognify - Your Learning Accelerator

Inspiration

Ever since I started working with AI APIs, I knew there had to be a better way to learn. As someone who's been a student most of my life, I'd been constantly juggling PDFs, lecture recordings, notes scattered everywhere - it was chaos. I wanted a single place where I could dump all my study materials and have AI help me actually understand them, not just store them. I'd record myself doing active recall sessions and wish I could analyze those recordings to see what I was missing. That's when I realized - why doesn't this exist yet? So I built it.

What it does

Cognify transforms how you study by becoming your personal AI learning accelerator. Upload your PDFs, audio recordings, lecture files - anything - and it creates personalized study content tailored to your learning level. It generates summaries, flashcards, practice quizzes, visual diagrams, and even finds relevant web resources. The platform analyzes your syllabus images to extract topics, creates different types of assessments (MCQs, short/long form quizzes, ultimate tests), and provides detailed feedback on your performance. It's like having a tutor who knows exactly how you learn best and creates content specifically for you.

How I built it

I built Cognify solo using React and TypeScript for a clean, responsive interface. The AI backbone combines Google Gemini for content generation and AssemblyAI for audio transcription. I implemented IndexedDB for offline storage so students can access their materials wherever they are. The architecture handles multiple content types - PDFs get parsed and analyzed, audio files get transcribed and processed, and everything gets organized into subjects and topics. I spent a considerable time fine-tuning the AI prompts to generate content that actually matches different learning levels, from high school foundations to professional development.

Challenges I ran into

The biggest challenge was making the AI responses actually useful rather than generic. Getting Gemini to understand the difference between explaining binary trees to a high schooler versus a computer science graduate took dozens of prompt iterations. Audio processing was another nightmare - handling different file formats, managing upload sizes, and ensuring transcription accuracy across various audio qualities. Managing state across all the different content types while keeping the UI responsive was complex. Rate limiting from APIs during testing meant implementing retry logic and graceful error handling. And honestly, designing an interface that doesn't overwhelm users with all these features was harder than the backend.

Accomplishments that I'm proud of

I'm most proud of how seamlessly everything integrates. You can upload a syllabus image, and it extracts all the topics automatically. Upload a PDF and audio recording of the same lecture, and it creates comprehensive study materials from both sources. The flashcard system that adapts based on what you know versus what needs review feels genuinely helpful. The quiz feedback isn't just "wrong", it explains why and suggests specific areas to focus on. Building a complete learning ecosystem in a week that actually works and feels polished is something I'm genuinely excited to show people.

What I learned

This project taught me that AI integration is as much about prompt engineering as it is about coding. I learned to think like an educator, not just a developer - understanding how people actually learn and retain information. Working with multiple APIs simultaneously showed me the importance of error handling and user feedback when things go wrong. I also discovered that the hardest part of building educational tools isn't the technology, but rather it's understanding the psychology of learning and translating that into features that actually help.

What's next for Cognify - Your Learning Accelerator

The immediate next step is adding collaborative features - study groups, shared flashcard decks, and peer quiz competitions. I want to implement spaced repetition algorithms to optimize review timing and add mobile responsiveness for studying on the go. Long-term, I'm planning advanced analytics to track learning patterns, and potentially voice based interactions for hands-free studying. The goal is to make Cognify the central hub for any serious learner, whether they're in school, preparing for exams, or advancing their career.

Built With

• assembly-api
• css3
• fetch-api
• file-api
• gemini
• git
• html5
• indexeddb
• javascript
• netlify
• npm
• react
• tailwindcss
• typescript
• vite