Prediction Copilot

Inspiration

The inspiration for PolyTerminal came from a desire to bridge the gap between traditional sophisticated trading interfaces (like Bloomberg Terminal or professional crypto exchanges) and the emerging world of prediction markets. While platforms like Polymarket have democratized access to information markets, their interfaces often lack the density and analytical depth that power users crave. We wanted to build a "Pro Mode" for prediction markets—a unified dashboard that combines real-time price action, on-chain whale tracking, and advanced technical analysis into a seamless experience.

What We Learned

Building PolyTerminal was a deep dive into the intersection of Game Theory, DeFi, and High-Frequency Data Visualization.

Market Microstructure: We learned how binary option pricing in prediction markets behaves differently from traditional assets, necessitating custom technical indicators (like modified RSI and Z-Scores) that account for the bounded $0-$1 range. Solana Architecture: Integrating heavily with Solana for on-chain tracking taught us the intricacies of the account model, PDA (Program Derived Address) derivation, and efficient RPC polling to track "whale" wallet movements in near real-time. Hybrid Data Fetching: We discovered that no single API was sufficient. We had to architect a hybrid layer that stitches together Polymarket's CLOB (Central Limit Order Book) for pricing, the Gamma API for metadata, and on-chain querying for positions, creating a unified data graph for the frontend.

How We Built It

The project is architected as a modern full-stack implementation:

1. Frontend (React + TypeScript + Vite) High-Performance UI: We used Recharts for bespoke, high-performance charting and pure CSS/Tailwind for a dark-mode, glassy aesthetic that feels premium and responsive. Custom Hooks Architecture: The data layer is abstracted into powerful hooks (useMarketHistory, usePythSignals , useWhaleMovements ) that handle caching, polling, and data normalization, keeping the UI components clean and declarative.
2. Backend (Node.js + Express) Aggregator Service: The backend acts as an intelligent proxy and aggregator. It polls multiple endpoints (Polymarket CLOB, Pyth Oracles, and On-chain RPCs) to synthesize "signals" like Volatility Anomalies or Whale Buys. Signal Engine: We implemented a logic layer that calculates technical indicators (MACD, RSI) on the fly, providing ready-to-consume signals to the frontend.
3. Solana Integration (Rust + Anchor) Whale Tracking Program: We wrote a custom Solana program (using Anchor framework) to track and register "whale" addresses. This allows the community to flag high-impact traders, whose moves are then monitored and displayed on the terminal.

Challenges We Faced

Data Synchronization: Synchronizing the "Current Price" across the chart, the order book, and the signal engine was difficult because different APIs update at different rates. We solved this by using the CLOB WebSocket as the single source of truth for real-time price updates while lazy-loading historical data. "Boring" Markets: Many prediction markets are flat for long periods. Implementing engaging visualizations for low-volatility assets was a UX challenge. We addressed this by highlighting relative changes and using Z-Scores to show statistical significance even in small moves. CORS and Proxying: Direct browser calls to certain API endpoints were blocked. Designing a lightweight backend proxy to handle these requests securely and efficiently was a necessary infrastructure hurdle. Conclusion PolyTerminal represents a step forward in how we visualize and interact with information markets. By treating outcome shares as serious financial assets with deep analytical needs, we hope to empower traders to make more informed decisions in the prediction economy.

Built With

• anchor
• react
• rust
• typescript