CrammAI

• 

  Landing Page

Inspiration

Back in university, I wasn’t limited by concepts — I was limited by time and information overload. I often thought, “If only I had an AI co-pilot that could summarize, quiz me, and keep me focused.” That thought became the foundation of CrammAI.

What it does

CrammAI is an AI-powered study co-pilot that:

Generates cram notes for fast learning. Creates mnemonics that stick. Acts as a 24/7 tutor for any uploaded material. Builds instant quizzes with smart grading via Gemini API. Offers live tutoring, where the AI explains and answers in real time.

How we built it

Frontend: Built with ReactJS + TypeScript for a clean UI. AI Layer: Powered by Google Gemini API for summarization, tutoring, and grading. Automation: Modular AI pipelines for notes, mnemonics, quizzes, and tutoring. Schema Enforcement: JSON schemas ensure reliable, predictable outputs.

Challenges we ran into

Handling different file formats (PDFs, notes, slides) while keeping the app responsive. Making AI outputs consistent and structured — solved with strict schema validation. Optimizing speed and performance for files under 10 MB.

Accomplishments that we're proud of

Built a production-ready web app that feels like a true study co-pilot. Designed automation pipelines that outperform single, monolithic prompts. Created a seamless user experience where students can learn faster and smarter.

What we learned

Prompt engineering is a skill: the more specific, the better. Micro-automations deliver more reliable results than mega-prompts. Balancing AI power with user experience is key to adoption.

What's next for CrammAI

Adding video/audio ingestion (lecture recordings, podcasts). Expanding conversational tutoring with more interactivity. Improving collaborative study features so groups can share notes and quizzes. Scaling to mobile for anywhere learning.

Built With

• dev
• docker
• google-cloud-ai-(deployment-&-cloud)-pandas
• google-cloud-ai-data-handling:-pandas-(for-preprocessing)
• google-gemini-api
• javascript
• modular-ai-pipelines-cloud-services:-vercel-(deployment)
• opencv-(data-&-testing)-git
• opencv-(for-testing-visual-inputs)-collaboration-&-dev-tools:-git
• power-automate
• power-automate-(for-task-scheduling)
• python-frontend-framework:-reactjs-ai/ml:-google-gemini-api
• python-reactjs-(frontend)-google-gemini-api-(ai-layer)-vercel
• react
• tensorflow-(for-experimentation)-automation-&-workflows:-json-schema-validation
• typescript