CalStreetBets

Inspiration

We wanted to make investing personal again. Most stock platforms are designed for institutions and experienced traders, cluttered, data-heavy, and intimidating for the everyday investor. Our inspiration came from the idea that financial tools should adapt to people, not the other way around. We wanted to give regular users the same analytical precision that professionals have, but with a focus on accessibility, customization, and clarity.

What it does and How we built it

Our app is a real-time, AI-powered stock tracker and portfolio dashboard that brings personalization to the center of investing. It allows users to build and manage their portfolios, view interactive charts, and analyze live market data all in one place. The key feature is its integration with Gemini, which uses user-defined constraints, such as risk tolerance, budget range, and sector focus, to deliver tailored investment insights and recommendations. The result is an intelligent, seamless experience that empowers users to explore, understand, and optimize their investments on their own terms.

We developed the frontend using Next.js and Tailwind CSS to create a clean, responsive interface that emphasizes simplicity and speed. On the backend, we used FastAPI connected to a MongoDB database to handle user data, preferences, and real-time portfolio updates. Gemini was integrated to interpret constraints and generate AI-powered investment suggestions that respond dynamically to user inputs. Together, this stack forms a cohesive system that bridges user intent, data visualization, and personalized insight.

Challenges & Accomplishments

Our first hurdle was finding free, reliable stock indicator APIs. We originally wanted to use a specific API offered by financialmodelingprep, only to discover that the “free” product was not as advertised. As not all of the group came with a solid background in finance, understanding the importance of the data to be included in CSB was a heavy topic of discussion among the group. Similarly, the technical aspects of FastAPI and Node.js configuration was a new concept for some of us that took time to fully grasp. Hard work pays off, and we have a lot to be proud of after the challenges we overcame. Our user-interface was a big accomplishment for our team as none of us have much prior experience in frontend. We managed to create a clean, modern and fluid UI while also being a crash course in React and TailwindCSS. Creating a smooth full-stack integration between Next.js, FastAPI, MongoDB, and Gemini, all within the hackathon timeframe with time to spare to consider future steps and room for improvement. We are also proud of our Retrieval-Augmented Generation, as it really elevated our real-time market data and context news from Polygon API. This makes the app’s recommendations both accurate and explainable. The integration of the APIs we used was very seamless and carefully thought out, and I hope that the user can see the careful planning behind CalStreetBets.

What's next for CalStreetBets

We’re expanding CalStreetBets into a truly personalized investing companion. The next major feature will be a stock news API that automatically curates articles and updates based on the user’s portfolio and behavioral data. Instead of showing generic headlines, the app will deliver relevant market news for the stocks part of the user portfolio and ones of potential interest complete with Gemini-generated summaries, sentiment classification (bullish, bearish, or neutral), and concise impact explanations tied to the user’s risk profile. Over time, the system will learn from user interactions to improve its filtering and recommendations. Additionally we would like to expand portfolio options to include other financial instruments such as ETF’s, commodities, or futures, potentially even including derivatives like options in later iterations.