WishWash

Landing Page
Chrome Extension Popup
Market Analysis
Time Horizon Divergence
Evidence Analyzer
ChatBot for the Market
User Portfolio Spread
Portfolio Aggregation

Inspiration

Prediction markets have a discovery problem. People read hundreds of articles about politics, sports, and tech news every day, but never connect that to trading opportunities on Polymarket.

We asked: What if Polymarket came to you?

Mission: Build the bridge between consuming information and taking action.

What it does

Chrome Extension

Detects what you're reading (Twitter, Bloomberg, Reuters) and surfaces relevant Polymarket markets in a popup. One click opens full analysis.

AI Analysis Platform

Scenario Generation: Analyzes real news and creates Bullish/Neutral/Bearish scenarios with probabilities
Time Horizons: Compare short-term (0-30 days) vs long-term (30+ days) to spot momentum vs fundamentals
Payoff Curves: TradFi-style visualization showing exact profit/loss at every outcome
Kelly Criterion: Optimal position sizing built-in
Portfolio Aggregation: Combined analysis across multiple positions

Voice Trading Advisor

Talk to AI!! Ask "Should I buy YES or NO?" and get instant, contextual advice.

How we built it

Frontend: React 19, Tailwind CSS, Recharts, Chrome Extension APIs, Web Speech API, LiveKit Client SDK

Backend: Node.js + Express, SQLite, Claude 3.5 Sonnet API, ElevenLabs API, LiveKit Server SDK, Polymarket APIs (Gamma, CLOB, Data)

Key Technical Wins

1. Time Horizon Architecture Different prompts for different timeframes:

Short-term: Breaking news, momentum, volatility
Long-term: Base rates, fundamentals, mean reversion

2. Kelly Criterion $$ \text{Kelly %} = \frac{p \times b - q}{b} $$ where $p$ = win probability, $b$ = odds received $\left(\frac{1}{\text{entry price}} - 1\right)$, $q = 1 - p$

3. Voice Context Injection Agent receives: market data, AI scenarios, evidence analysis, user portfolio—allowing truly contextual responses like "Based on the Bullish scenario, a $100 YES position would net $376 profit."

Challenges we ran into

Voice Latency

Early tests had 3-4 second delays. Solution: Hybrid approach (browser speech recognition → backend → ElevenLabs TTS) reduced latency to ~1 second.

Accomplishments that we're proud of

Mathematical correctness - proper Kelly Criterion, accurate P&L for YES/NO

Novel AI - time horizon divergence actually useful, not hype

Natural voice - first voice trading advisor that sounds human

Professional visualization - TradFi-quality curves

What we learned

Technical: Binary options math, LiveKit architecture, AI prompt engineering, Chrome extension development

Voice insight: Voice doesn't replace analysis; it makes analysis accessible to people who can't interpret charts

What's next for WishWash

Vision: WishWash becomes the default way people discover prediction markets. Like Honey shows coupons while shopping, we show betting opportunities while reading news.

Prediction markets become accessible to everyone, not just quants. That's the world we're building.