GoldLens

Our Chrome extension brings a built-in chatbot and class and professor information directly into the registration website, helping students make informed decisions without leaving the page.

Comment

Inspiration

Our inspiration came from the inefficiency of managing multiple tabs to research courses and professors, so we wanted to streamline the entire experience into a single interface.

What it does

GOLDLens is a Chrome extension that overlays real-time course intelligence directly on UCSB's GOLD registration system. When you hover over any course, it displays:

  • Historical grade distributions (A/B/C/D/F percentages across all instructors)
  • Professor ratings and reviews from RateMyProfessor, including difficulty scores and student comments
  • Current quarter instructors highlighted with stars, fetched live from the UCSB API
  • Per-instructor grading patterns so you can compare how different professors grade the same course

We also built an AI-powered chatbot that opens in a side panel. It's context-aware so when you click the search icon on a course card, the chatbot automatically loads all the grade data, RMP reviews, and course info, then lets you ask questions.

How we built it

  • Frontend: Vanilla JavaScript Chrome Extension (Manifest V3) with Shadow DOM for isolated UI components
  • Backend: FastAPI server that handles course data bundling and AI interactions
  • AI: Google Gemini SDK for generating course insights and powering the chatbot
  • Data Sources: UCSB API for live course/instructor data, scraped RateMyProfessor data using Playwright, and compiled historical grade distributions

Challenges we ran into

  • Data integration: Merging grade data, RMP reviews, and live UCSB API data into a unified bundle that the AI could reason about coherently
    • RAG tuning: Getting Gemini to pull course-specific data accurately without hallucinating or giving generic responses—we had to craft strict prompts that force data-driven answers

Accomplishments that we're proud of

  • Seamless UX that feels native to GOLD
  • Context-aware chatbot that knows exactly which course you're asking about
  • Real-time instructor detection that highlights who's teaching this quarter

What we learned

  • How to build Chrome extensions with Shadow DOM for style isolation
  • Integrating multiple data sources into a cohesive RAG pipeline
  • Prompt engineering to get consistent, structured JSON outputs from LLMs
  • Async scraping at scale with Playwright and rate limiting

What's next for GoldLens

  • Expand to support schedule planning and workload balancing across multiple courses
  • Deploy the backend so it works without running locally

Built With

Share this project:

Updates