GLWM (Get Locked-in With Me)

Chart summary example
Personal setting example
Visualization example

Inspiration

Sometimes, we're faced with long and complex text for studying, research, or work, and it’s easy to lose focus or miss the key ideas. We wondered, what if reading online could feel lighter, clearer, and more interactive? And why copy and paste materials to your online notebook app when you can do the same thing on the website? Those thought inspired us to build a tool that adapts to different learning styles, whether you’re a visual learner, a student, or someone who prefers quick key-point summaries.

What it does

Our Chrome extension allows users to select any text on a website and instantly generate: ✅ A concise summary using Google Chrome’s on-device AI ✅ A visual representation (e.g., diagrams, flowcharts, charts) when relevant ✅ A highlight-based note-taking system with comments ✅ A personal “Summary Style” based on user settings (e.g., student, researcher, developer)

It turns reading into a faster, smarter, and more personalized experience. It's like a virtual notebook!

How we built it

Chrome’s Summarizer API for key-point summaries
Chrome’s Prompt API for prose-style TL;DR summaries
Custom DOM parsing to extract: Nearby visual elements like tables, figures, and charts
Commenting and annotation storage using Chrome local storage
Rendered Mermaid.js diagrams for AI-generated visuals

Challenges we ran into

Getting the Summarizer API and Prompt API to provide the wanted summary
Handling complex DOM structures on different websites (news, research papers, Wikipedia, etc.)
Avoiding conflicts when extracting selected text vs. full-page text
Getting Mermaid diagrams to reliably generate valid syntax from natural language text

Accomplishments that we're proud of

Runs entirely on-device, which keeps user data private
Generates both text summaries and visual diagrams
Includes a highlight + commenting system for active learning
Works across a wide range of web pages

What we learned

Learned how to design and optimize browser extensions for performance and stability.
Developed strategies for blending on-device processing with cloud-based AI for improved results.
Discovered the importance of user-centric feedback loops when managing stored data and context.
Explored ways to preserve context so AI summaries stay accurate and relevant to what the user sees.