How we got here

The project used to be called ZkVanguard. People kept asking if we were a bond fund. It sounded like your dad’s retirement plan. So we renamed it to ZKward. Same idea, slightly awkward to say out loud. You’ll get used to it.

What it is, in one sentence

A savings pool that trades for you and shows its math, so you don’t have to take our word for it.

How it works, plainly

You put in USDC. A small crew of computer programs reads what people are betting on in prediction markets (like Polymarket) and lines up trades on crypto based on that. When it makes a trade, it also produces a little cryptographic receipt that anyone can check. The receipt proves the trade followed the rules — without revealing your position.

Think of it as an autopilot with a black box flight recorder. Except the recorder is public, and the autopilot won’t take off in bad weather.

Where it started

A hackathon weekend. Too much coffee. One question: if computers move real money, why does anyone trust their reasoning? We thought we’d answer in two days. It took eighteen months.

Along the way we won five hackathons, contributed to Tether’s wallet toolkit, and shipped on a handful of blockchains. Most of the test deployments lasted about an afternoon. SUI stuck — it’s fast, cheap, and doesn’t break when you look at it wrong. That’s where the money lives now.

What we ship today

• A real pool on SUI mainnet. Small on purpose — the contract itself caps deposits at $10,000. We wanted to prove it works before scaling. Slow is the point.
• Eight safety switches that can pause the whole thing if something looks off. One tripped yesterday, after a rough day of testing (more on that below).
• A learning loop: the AI grades its own trades, retrains overnight, and tries again. Sometimes it gets better. Sometimes it finds new ways to fail. We publish both.
• A cryptographic receipt for every trade. Not a screenshot. A proof.

Why we tell you when it loses

Most projects only talk about their wins. Right now, our test trader is down about $68,000 on paper. We’re telling you because that’s the whole point. If you can’t see the losses, the wins don’t mean much.

Also, hiding a loss is how a small loss turns into a much larger one three months later. We tried that once. Do not recommend.

Who’s making it

Mostly one person, working from nine time zones away from most of you. Computer science degree, some years at big companies, an unreasonable number of side quests. Funded by savings and small grants that pay out roughly around the time the founder starts skipping lunch.

The paper trader has occasionally cost more per month than the founder eats. We consider this on-brand.

Where this goes next

Two things need to happen. First, prove the trader stops losing money in the test pool. Then raise the deposit cap, one careful step at a time, with a real drawdown test at each step. If we skip a step, please yell at us on Telegram.