The AI-native stack for prediction markets.

There's a weird thing happening right now that most people haven't noticed.

When the 2024 election was too close for every pollster to call, prediction markets had it right. When weather models disagree on whether tomorrow will hit 90 degrees in Miami, there's a market where people are betting real money on the answer. And they're weirdly accurate.

This isn't a coincidence. When you force people to back their opinions with money, the noise disappears. The guy who "just has a feeling" bets small or stays home. The person who actually did the work bets big. The price that comes out of that process is something we don't really have a word for yet. It's not a poll. It's not a forecast. It's closer to the truth than either.

We think this is one of the most important developments in how information works, and almost nobody is building serious infrastructure around it. The data is scattered. The tools are primitive. If you want to study how these markets form prices, where they're wrong, where they're early, you're basically starting from scratch every time.

We're building the research stack that should already exist. Not a trading app. Not another exchange. The infrastructure layer underneath. The part that takes raw prediction market data and makes it actually useful for people doing real work.

But we also think about something that doesn't get talked about enough in this space.

Prediction markets are new. The regulation is still being written. And the truth is, not every market that can be created should be. You can create a market on almost anything. That doesn't mean you should. When markets are designed around tragedy or engineered for outrage, they don't prove that prediction markets work. They prove that anything can be gamified. And that makes it harder for the markets that actually matter to be taken seriously.

We chose to focus on outcomes you can verify. Weather. Sports. Economics. Things that resolve against real data, not opinion. Not because the other stuff isn't profitable. Because we think the builders who show up now are deciding what prediction markets become. And we'd rather they become trusted infrastructure than a punchline.

We use these markets ourselves, every day. We build what we need. Increasingly, the researcher is an AI agent. We built for that.

The foundations of this space are being built right now. We'd rather build them right.