RobbingHood

Inspiration

Finance content is everywhere, and apparently everything is a signal if you stare at it long enough. So we built RobbingHood, a slightly unhinged Chrome extension that assumes whatever you’re reading contains hidden “alpha.” It reads the page, jumps to conclusions, and delivers a confident investment take—because why let fundamentals get in the way of vibes? (For entertainment only, obviously.)

What it does

RobbingHood extracts readable text from the current webpage, sends it to an AI backend, and returns a stock or crypto recommendation with a clear explanation and confidence level. For example, an article about heavy rain might result in a recommendation like $UBER, based on increased ride-hailing demand. The extension also includes a Vibe Check feature for rapid re-analysis and a Demo mode to test the system without loading a live webpage.

How we built it

The frontend is a Chrome extension built with React, TypeScript, and Vite, using Manifest V3 and a Side Panel UI. A content script handles webpage text extraction and communicates with the backend, while Tailwind CSS is used for styling. The backend is implemented with Python and FastAPI, using the OpenAI API to generate structured recommendations and Yahoo Finance to retrieve live market data.

Challenges we ran into

Scraping clean, relevant text from arbitrary webpages is difficult due to varied layouts, dynamic content, and irrelevant noise. Integrating Manifest V3 components and the side panel required careful coordination. Ensuring consistent, structured AI outputs required prompt iteration, and setting up reliable communication between the extension and a local backend added additional complexity.

Accomplishments that we're proud of

We built an end-to-end Chrome extension that can analyze any webpage and generate explainable, confidence-scored investment recommendations. The system integrates live market data to provide real ticker context and includes a clear project structure with a straightforward local setup process.

What we learned

We gained practical experience building Chrome extensions with Manifest V3, designing a simple service architecture, and applying prompt engineering to enforce structured AI outputs. The project also reinforced the importance of early integration and tight scope management under hackathon time constraints.