Quillin'
Ask tutor-bots questions directly from your Apple Pencil notebook
Role
Product Design
Front-end
Tools
Figma
SwiftUI
Railway
FastAPI
Platform
iPadOS (SwiftUI)
Tags
UX of AI
SwiftUI
iOS
Overview
At Columbia DevFest 2025, my team developed a handwriting notes app inspired by the challenges of keeping up with handwritten notes in class and the desire for better cross-platform sharing. Combining handwriting recognition with an LLM tutor bot, the app allows users to take notes and interact with an AI assistant for clarification or expansion.
Starting from a Figma design, we built an iPadOS app in Swift with a FastAPI backend. Third-party services—including Groq, Google Cloud, and Railway—supported handwriting recognition and cloud functionality.
Pain Points
In the user's shoes
User Story
Here are some user stories we drafted, thinking about what kinds of use cases we could build for.
As a student in a large lecture, I want to ask a question about a concept I don't understand anonymously, so that I can get clarification without feeling intimidated in front of 300 people. |
As a student trying to follow a lecture, I want to quickly look up a term I just heard (like 'linear dependence'), so that I can understand the concept in real-time without falling behind or missing the next part of the lecture. |
As a student on a small lecture desk, I want to take notes and search for information on a single device, so that I don't have to juggle a laptop, notebook, and pen in a cramped space. |
As an engineering student working on homework, I want to take a picture of my handwritten notes or a problem, so that I can get an explanation of symbols and letters that I don't know how to type. |
Solutions
SOLUTION 1
Transcription to Latex
For notes where you’re unsure what you wrote, quickly convert handwriting into readable and shareable math notation.
SOLUTION 2
One-step to asking an LLM
Highlight your notes and send to an LLM to get answers and explanations.
SOLUTION 3
Suggested questions
The app proposes relevant questions automatically to guide your learning, so that the user doesn't need to switch to typing on their iPad to ask an LLM chatbot.