Canvas Wrapped

Home Page
Persona and AI Personalized Description
Classes Average Score + Advices from AI
Vibe Check + AI Roast
That's a Wrap!
How to Get Your Canvas Token
Privacy & Security

Inspiration

Inspired by Spotify Wrapped, we created our own "Canvas Wrapped," a fun and personalized way (using AI) for students to reflect on their academic habits through canvas data. Instead of showing grades, it transforms everyday school data into engaging and fun insights, like your record for latest assignment submission to procrastination patterns. Canvas Wrapped is meant to be engaging, meaningful, and social media trend-worthy.

What it does

Canvas Wrapped transforms students' terrifying Canvas data into a fun, personalized recap of their academic habits and experience, highlighting patterns like assignment submission timing + patterns, procrastination trends, and personalized analysis using K2 Think V2 model to generate meaningful academic personalities with sustained reasoning across hundreds of data points simultaneously.

Challenges we ran into

One of the main challenges we ran into was accessing Canvas data through the Canvas API, as it was difficult to retrieve complete and accurate datasets, since certain information, such as courses under "completed" flag, was not easily accessible on the first try.

Another notable challenge we faced was designing the data pipeline from end to end, specifically routing Canvas API data into our own computed analytical metrics and combining them with LLM-generated insights. Structuring this flow (through routings) so that raw data, mathematical analysis, and AI outputs could be consistently integrated and then passed to the frontend for interactive visualizations required multiple iterations to properly align the backend logic with the UI experience.

Accomplishments that we're proud of

We successfully built a data pipeline that allows users to input their Canvas token and instantly retrieve and process their Canvas data. This data is then efficiently routed through our system into the LLM while carefully managing context limits to avoid overloading the model. Within seconds of token submission, the processed insights are generated and directly translated into an interactive UI experience, creating a fast and seamless end-to-end flow from raw data to user-facing visualization.

What we learned

We learned the importance of cutting down scope and focusing only on what's truly necessary for the product. Under time pressure, we realized that trying to build too many features can slow progress, so streamlining our ideas helped us stay aligned and move faster. This experience also taught us how to prioritize core functionality and make clearer decisions throughout the development process to efficiently build a working product within a limited timeframe.

What's next for Canvas Wrapped

We want to add an AI chat bot that users can talk to and receive advice on courses, things like what courses they should take in the future and how well they will do if they take difficult classes. The chat bot will analyze their data and predict whether the user will do well in the classes they take, factoring in things like their strengths, how well they do in different courses, their time management and how well they perform in exams and finals.