Project Info

shortcutz

Devpost

This project did not submit a demo video on Devpost.

Inspiration

Aman and Bhargavi spent the winter quarter being teacher assistants for local high school math classes as UCSC Calteach interns. There, they were shocked to find out how many students had a lack of math foundational topics, to the point where they used calculators for basic addition because they did not know how to do it themselves. It was heartbreaking watching the students feel inadequate because they could not learn advance topics due to their lack of foundational skills. The students were not deterred by this challenge, and showed up to class everyday to get a little better. There, Bhargavi and Aman learned that the students picked up math concepts much faster when they were taught a "shortcut" (I.E. 5 times 10 equals 5 following a 0 or 50). Relearning early math concepts that were reinforced by these "shortcuts" ensured they remembered the lessons they learned. That's how our group came up with the app idea "Shortcutz". We built it for the students that were let down in their earlier math education but are determined to continue learning. Having their growth reinforced by the shortcuts and a fun layout is the main goal of our app.

What it does

Shortcutz is an AI-powered math training app that dynamically adjusts difficulty based on student performance. It challenges users with multiplication and division problems, provides instant feedback, and delivers personalized AI-generated hints and encouragement powered by Gemini AI.

How we built it

We built the app using React.js for the frontend, leveraging hooks like useState, useEffect, and useRef to manage the quiz state, user inputs, and adaptive difficulty. The smart feedback loop comes from integrating Gemini AI via the Google Generative Language API, which takes the question and user response context to generate helpful and friendly feedback. The difficulty adjusts in real time using a performance-based algorithm that evaluates accuracy and speed.

Challenges we ran into

API integration: Parsing the Gemini API’s nested response format correctly took a little tinkering. Adaptive difficulty: Balancing the level progression logic to feel fair, yet challenging, was a puzzle of its own. Layout organization: Restructuring the UI to make the app feel clean and intuitive, especially with split panels for questions vs. feedback.

Accomplishments we're proud of

Seamless integration of LLM-based feedback into a math app — a super cool AI/education combo. A smooth user experience that tracks performance and improves dynamically over time. Created a complete tool that is both helpful for learners and satisfying to build!

What we learned

How to connect and manage LLM APIs in a React environment How to craft meaningful, structured prompts that lead to actionable AI responses The power of thoughtful UI/UX when combined with smart algorithms — clarity = confidence in learning!

What's next

Add leaderboards and streaks to gamify practice Integrate voice input and read-aloud questions for accessibility Expand question types to cover fractions, algebra, and word problems Store user history in the cloud for long-term progress tracking Add multi-language support and mobile responsiveness

Analysis

Compare with all teams

View

Metric

Figures cover GitHub contributors during the hackathon window. A co-authored commit counts in full for each author, so per-member totals add up to more than the whole-team figures.

Technology

Found in codeClaimed only
  • CSSIn code
  • HTMLIn code
  • JavaScriptIn code
  • ReactIn code
  • Google GeminiClaimed

4 of 5 appear in the indexed code. 1 claimed on Devpost could not be matched to code, which may simply mean the tool leaves no trace in the repository.

AI coding agents

No AI coding agent signals were found in this repository.

Detected from committed agent config files and commit authorship. Absence of a signal is not proof an agent was unused.

Codebase size

Source size

18 KB

Source files

9

Counts recognized source files only; vendored directories, binaries and lockfiles are excluded, so this is smaller than the repository on disk.

0 stars