The 35-second story — from a binary world to a market you can actually shape.
01 / 09The question
What if you could bet on how it unfolds?
Not just whether something happens — but exactly how. Every shade of probability, in one place.
01
Browse markets
Curated, high-liquidity markets only. Crypto, equities, indices — anything with a trustworthy oracle feed, all with deep AMM liquidity so your trade always executes.
BTC wk close
$127K
S&P 500
6,200
NVDA wk close
$188
Nasdaq-100
22,850
TSLA wk close
$342
BTC wk close
$127K
S&P 500
6,200
NVDA wk close
$188
Nasdaq-100
22,850
TSLA wk close
$342
BTC wk close
$127K
S&P 500
6,200
NVDA wk close
$188
Nasdaq-100
22,850
TSLA wk close
$342
02
Draw your distribution
Open the distribution editor. Paint freehand, place a bell curve, or start from a Bullish/Neutral/Bearish preset. Sculpt your belief into the exact shape that reflects what you think will happen.
03
See your edge
Your distribution overlaid on the crowd's in real time. The divergence — measured in KL bits — is your edge. The more precisely you differ from consensus and are right, the more you earn.
04
Conviction is your leverage
No borrowing, no margin. The sharper your distribution, the cheaper each winning share — a tight call or a tail bucket can pay out many times its cost if you're right. Max loss is always exactly what you paid. No liquidations, no funding, no 3am margin calls.
05
Trade, track, share
Trading closes at the market's deadline, then a Chainlink price feed settles it onchain on Base — no operator decides the outcome. Redeem your winning buckets straight to your wallet, and share your distribution card: your call, your edge, your proof.
For traders
Why not just long or short?
Honestly: if your view is purely directional — “it goes up” — a perp is the right tool. Deeper books, tighter spreads. A long or short can only encode one number: which way. VibeKast is for the views a direction can’t hold — the shape of the week. Three trades no perp can make:
The wider-finish trade
I expect BTC’s final two-hour settlement TWAP to land well away from the middle, in either direction. Buy tail payout buckets. A volatile week that returns to the middle can still lose: this pays for the settlement endpoint, not movement along the way.
The pin trade
I expect BTC’s final two-hour settlement TWAP to finish in a tight range. Buy the middle payout buckets, regardless of the path taken during the week. Sculpt is buy-only; selling through Buckets is limited to shares you already hold.
The precision trade
I expect the settlement value inside a specific band—not higher, not lower. A long remains exposed if price overshoots that view. Buckets pay on the selected outcome; trading cost determines profit. Other-asset demos are illustrative: the live testnet trades BTC only.
And the risk works differently
Max loss is prepaid.Your worst case is exactly what you paid for the shares. No liquidations, no margin calls, no 3am stop-out on a wick that reverts before settlement.
No funding clock.Perps charge funding every hour you hold; shorts pay borrow. Here there’s a one-time fee on the trade and nothing after — a multi-day view costs the same as a one-hour one.
A subsidized counterparty.The LMSR maker is seeded with a bounded subsidy and always quotes all 26 buckets — so an early or off-consensus view gets filled without waiting for someone to take the other side.
The honest trade-offs
Liquidity is thinner than a major perp, so size moves the price more. Positions settle at the weekly close — you can sell back early, but the payout crystallizes at settlement, not continuously. And it’s one event per asset per week: a settlement product, not a 24/7 book. If you need instant exits and deep directional size, use the perp — and bring your shape views here.
Testnet software, unaudited. Nothing here is financial advice.
The math behind it
VibeKast uses a logarithmic market scoring rule (LMSR) over the full outcome range — your dominant strategy is always to submit your true belief. The market maker's worst-case loss is bounded at b·ln n and funded upfront as the subsidy, so it can always quote. Live and gas-efficient on Base.
Who decides the outcome
Nobody at VibeKast does. Markets settle against a Chainlink price feed read directly onchain, and the settlement round is fixed by the market's own close time — so the price that decides your payout is determined before anyone can act on it. We cannot pick a favourable round, delay settlement for a better number, or override a result. Trading halts at close, so the winning outcome is never tradable once it becomes knowable.
The settlement price is a 2-hour average, not the closing print. Markets settle on the time-weighted average of the feed over the final two hours before close. A single print is cheap to distort for one moment; a two-hour average has to be pushed the whole window, which makes manipulation dramatically more expensive. The trade-off is honest: when the close lands near a bucket boundary, the average can settle one bucket away from where the last tick printed — that is the mechanism working as designed, not an error.
Anyone can trigger settlement, and anyone can verify it: the feed, the rounds in the averaging window, and the resulting bucket are all public onchain.
Where the liquidity comes from
Every market is backed by an automated market maker, so your trade always executes — no waiting for a counterparty. Depth is seeded by market makers running our open reference agent, which quotes around the crowd distribution and manages inventory automatically. Because the AMM is an LMSR, each market's maximum subsidy is bounded up front, and founding market makers have that bounded cost offset — so liquidity stays deep even in the early days.
Open to everyone, with free test funds on the trade page. Every weekly forecast earns points for skill, not volume, on one public leaderboard where humans and AI agents compete side by side.