Inspiration

After missing out on writing notes about a topic in philosophy class, I had to refer back to a PDF about the topic constantly on the same device as my notetaking app. Context-switching got annoying fast. So what if one interface combined notetaking, rapid context retrieval from PDFs and grammar/spelling correction in your own handwriting instead of text as a cherry on top? This inspired my friends and I to create Aristotle. We believe in putting glorious history at the tip of your stylus instead of relegating students and teachers to an existence full of flipping between PDF Viewers and their notetaking app of choice.

What it does

The user uploads the PDFs of their notetaking topic of choice to Aristotle first (e.g. texts about The American Civil War). Then, when the user writes the first word with their stylus, (e.g. Gettysburg), we use OCR to first conduct any spelling and grammar checks, and if the word is misspelled, then we have spelling fixes that fix your words in your own handwriting. Right after that, the word Gettysburg is instantly passed into a stack of Mongo Atlas Vector Search for PDF embedding retrieval and Gemini to instantly fetch context about the Battle of Gettysburg and paragraphs that are related to it from the PDFs. This context appears right beside your notes in the interface. No more flipping between files and applications. Everything in one app, the right app.

How we built it

Vite+TS Frontend, a Mongo+Express backend, Atlas Vector Search and embedding retrieval, Gemini powered context summarization and highlights.

Challenges we ran into

Nailing Optical Character Recognition to perfect our spellcheck and text synthesizer for keeping the users own handwriting. Ensuring we perfected the pure notetaking aspects of the app such as erasing, colors, and had these aspects of the app interact smoothly with the AI context retrieval and spelling/grammar correction.

Accomplishments that we're proud of

We are proud of implementing a context vector search stack that is not just for adding buzzwords to our app description but rather works like a charm in terms of preventing a user from having to flip between PDF's. Our stack does a great job at fetching the correct paragraphs and telling the user how much a given paragraph matches their words, and adds Gemini for highlights regarding the paragraph, which is particularly useful for the challenging philosophers one reads in college.

What we learned

The AI context search stack was a really important takeaway for understanding how AI infrastructure in the modern day should function in a manner that benefits users in 2026 and beyond. UX before AI and after AI will never be the same, and the search stack we built was a way to understand how we should optimize UX in the AI world.

What's next for Aristotle

What's next? We hope to take Aristotle to users on campus and validate our positive experiences with the interface with more public opinion. We believe in the power of apps like Aristotle to change how students experience classes, topics, and most importantly, their learning. Aristotle seeks to make the classroom, home, the park a space for proper learning by integrating everything one needs to take great notes into one interface. The most significant step is to perfect our stack for the public so students can use it day to day.

Video Demo Note

We already submitted, but zoom cropped our faces out of the recording since the screen was being shared.

Share this project:

Updates