Finley

Managing your investments can be really hard and time consuming. That's why I've built Finley, a Gemini AI companion that syncs with your financial portfolio to turn raw data into a clean stategy.

Comment

Inspiration

My inspiration for Finely came from a very personalised realisation; managing personal investments can feel incredibly complicated and overwhelming. Millions of retail investors track their money using handmade spreadsheets, but they lack a dedicated financial advisor to explain how market shifts can impact their porfolio. I wanted to build an intelligent companion that turns intimidating market noise into empowering conversations.

What it does

Finely is a warm financial companion, for everyday retail investors. Instead of overwhelming the investors with complex knowledge and intimidating spreadhseets, Finely takes a users portfolio and translates it into an action plan on how to proceed. It explains the stock market using real world analogies and is able to discuss investing safely.

How we built it

This app features a responsive React, TypeScript and Tailwind CSS frontend styled with inviting emerald and sand tones. The backend is run on a Node.js server, leveraging the new @google.genai SDK to integrate Gemini 3.5 flash with strict JSON responseSchema constraints for structured dashboard rendering.

Challenges we ran into

My biggest hurdle was synchronising my pipeline against the SDK ecosystem. I repeatedly hit backend failures when older documentation pointed me towards deprecated model configurations; requiring me to start from scratch. I also had to implement dynamic routing to avoid Cross-Origin Resource Sharing (CORS) error through replit's public web proxies.

Accomplishments that we're proud of

I successfully created an empathetic AI persona that feels like talking to a friend over a cup of coffee rather than navigating a spreadsheet full of cold knowledge. Moreover, I am proud of building a robust full-stack structure that cleanly passes dialogue down to the SDK layer without ever lagging or breaking client UI.

What we learned

Building this under a tight deadline taught me just how fast AI tools are changing- one minute a model is standard and the next it becomes deprecated; making backend configurations crucial to keep the app live and working. I also gained massive hands on experience programming and designing the app. I learnt how to catch and rectify tiny bugs and configure the frontend developments.

What's next for Finley

Next I want to expand Finely's reach to support global markets- specifically tracking Indian NSE equities and local UAE government funds. I also plan to build interactive charts that can map out growth trends alongside Finely's conversational breakdowns, and create an automated file parser so users can upload their personal portfolio spreadsheets.

Built With

Share this project:

Updates