Inspiration

Paid betting communities on Discord and X are filled with analysts and channel owners claiming impressive win rates and massive profits. However, there is one major problem: nobody independently verifies their track records. Analysts can highlight winning bets, delete losing predictions, or advertise misleading success rates.

We wanted to bring transparency and accountability to this space. That inspired us to create PickProof, a Chrome extension that helps users distinguish genuinely successful analysts from those who only appear successful.

What It Does

PickProof tracks both verified and unverified betting analysts and channel owners across platforms like Discord and X.

  • Verified analysts: Analysts who voluntarily register with PickProof.
  • Unverified analysts: Analysts who have not registered but whose publicly accessible picks can still be tracked.

PickProof analyzes available picks from the past 30 days, extracts betting information from screenshots, and compares predictions against actual Gemini prediction market settlements.

It ranks analysts using three key metrics:

  • Win Rate: Percentage of successful predictions.
  • Net Units: Actual betting profitability.
  • ROI: Return on investment across tracked picks.

For live predictions, PickProof also displays the corresponding Gemini prediction market contract and current price, helping users evaluate opportunities directly.

How We Built It

We designed PickProof as a Chrome extension that connects social betting content with verifiable prediction market outcomes.

Our technology includes:

  • Chrome Extension: Provides an accessible interface for tracking and comparing analysts.
  • Tesseract.js: Performs free, browser-based optical character recognition (OCR) to extract predictions from betting screenshots.
  • Gemini Prediction Markets: Provides real market settlement outcomes and matching contracts.
  • Performance Analytics: Calculates win rates, net units, and ROI to create transparent analyst rankings.

By processing screenshots locally with Tesseract.js, we eliminate per-image OCR API costs and make tracking more scalable.

Challenges We Faced

One of our biggest technical challenges was extracting accurate betting information from screenshots with different layouts, formats, and image quality.

Another challenge was matching predictions made on social platforms with the correct Gemini prediction market contracts. Different naming conventions and market structures made accurate matching especially important.

We also needed to distinguish registered analysts from unregistered analysts without excluding publicly available predictions, while ensuring that performance metrics were based on actual outcomes rather than advertised results.

What We Learned

Building PickProof taught us how to combine browser extensions, OCR, prediction market data, and performance analytics into one consumer-focused product.

More importantly, we learned that a high win rate does not necessarily mean profitability. An analyst can win 60% of their bets and still lose money depending on the odds and position sizes.

This reinforced why transparency requires more than simply counting wins and losses.

What's Next

Our vision is to make PickProof a trusted source for evaluating betting analysts across social platforms.

We want to expand platform coverage, improve prediction extraction and market matching, introduce longer-term performance histories, and help users make more informed decisions using verified historical data.

Our goal is simple: Stop trusting screenshots. Start trusting proof.

Built With

Share this project:

Updates