Phitien

AI Market Generator and Market Filters
Open Positions

Inspiration

We wanted to build a prediction market experience that feels fast, understandable, and fun, while also showing how AI can help create relevant markets from real-world news. Phitien was inspired by platforms like Polymarket and Kalshi, but aimed at a simpler experience with instant interaction and a cleaner flow.

What it does

Phitien is a prediction market web app where users can browse markets, buy YES or NO positions, track their balance, and view their open bets. It also includes an AI-powered market generator that creates new prediction markets from current topics, making the platform feel dynamic instead of relying only on hardcoded data.

How we built it

We built Phitien with a React frontend and a Node.js/Express backend. The frontend handles the market board, category filtering, betting interface, balance display, and positions list. The backend exposes API routes for markets, bets, and funds, and stores data in local JSON files for a lightweight prototype setup. We also integrated an AI workflow that can generate new markets based on live topics and append them into the app.

Challenges we ran into

One challenge was keeping the frontend and backend state in sync after actions like placing bets or generating markets. Another was structuring the project so the UI stayed responsive while still simulating realistic market behavior like probability shifts and volume changes.

Accomplishments that we're proud of

We’re proud that we turned the idea into a working end-to-end product that feels usable instead of just a mockup. The market cards, betting flow, category filtering, balance updates, and open positions all work together in a coherent experience. The AI market generation feature also gives the project a strong differentiator by making the app feel timely and extensible.

What we learned

A big takeaway was managing state between the frontend and backend so that bets, balances, and market probabilities all update correctly. We learned a lot about coordinating frontend state with backend updates, especially in a betting-style app where every action changes multiple parts of the UI.

What's next for Phitien

Next, we want to replace local JSON storage with a real database and add authentication so multiple users can have separate accounts and portfolios. We also want to improve market resolution, add admin controls, make AI-generated markets more reliable, and introduce richer analytics like price history, leaderboards, and portfolio performance over time.