Inspiration
When we research/open tabs on the web, we open so many tabs that their titles are not clearly visible after some point, and we cannot tell which tab is which. We forget what each page said, we move to pages that are not related to our work, and we lose everything when the tabs are closed. Sharing our research with a teammate is also difficult.
What it does
Research Map saves every page you read and shows it on one map. Gemini writes a short summary of each page and groups pages on the same topic. If you set a goal, the app tells you when a page is not related to it. At the end, you get a report with all your sources, and you can share your map with your team through a link.
How we built it
We built the web app with Next.js and the server with Go. The AI part uses Python and the Google Gemini API. A small browser extension saves the pages you read. The data is stored in PostgreSQL, and the app runs on Firebase Hosting and Google Cloud Run.
Challenges we ran into
The free Gemini limits were reached quickly, so we added extra keys and a second model to keep the app working. Our goal check also gave wrong warnings at first, so we made it warn only when a page is clearly unrelated.
What we learned
We learned to plan the problem, the user, and the result before writing code. We also learned how to keep an AI app working when the AI is busy.
What's next for Research Map
We plan to publish the extension on the Chrome Web Store, let users ask questions about their saved pages, and add support for Hindi and other Indian languages.
Log in or sign up for Devpost to join the conversation.