Field Guide
What this is, what it shows, and how to use it. Five minutes, no jargon.
What gap133 is
gap133 is a cross-venue prediction market terminal. It watches two of the largest prediction markets in the world, Polymarket and Kalshi, at the same time, and shows you one number most traders never see clearly: the price gap on the same event between the two venues.
Prediction markets are exchanges where people trade on the outcomes of real events: elections, Fed decisions, sports, crypto prices. Prices are quoted in cents, and a price of 42¢ means the market believes there is a 42 percent chance of that outcome. Polymarket and Kalshi are the two biggest venues. They are open to different audiences, run on different infrastructure, and are priced by different crowds. So the same real-world event often trades at two different prices at the same moment.
That difference is the gap. gap133 finds it, measures it in cents, then does the two things other scanners skip: it subtracts the taker fees so you see the net edge, and it reads both order books so you see the size you could actually put on. We rescan the whole board every few minutes, and we publish our matching evidence so you can check the checker.
What the terminal shows
Every row on the board is one prediction market matched across both venues:
- Event: the question being traded, with the Kalshi market ID underneath.
- Kalshi yes: the current YES price on Kalshi, in cents.
- Polymarket yes: the current YES price on Polymarket, in cents.
- Gap: the edge at the touch, in cents, before fees: buy one venue at its ask, sell the other at its bid. The bar under the figure shows relative width at a glance.
- Quote gap: the small line under a gap when it differs by half a cent or more. It is the plain Kalshi bid vs Polymarket price, shown for context; the charts track this number. Inside spread means there is no executable edge right now, so the row is dimmed and never alerts.
- Net: the edge at the touch minus both venues' taker fees at the prices you would actually trade. The net + badge appears only when that survives real costs on a pair whose settlement terms match closely.
- Size: how many contracts you could trade on both venues at the touch (the smaller of the two order-book legs). Each contract pays $1 at settlement, so both legs together tie up roughly $1 per contract. A 3¢ edge on 400 contracts is about $12; on 40,000 it is a trade. Open a row to see what each leg costs.
- Live: both venues printed a trade within the last 24 hours (green both live), or one side has gone quiet and the row is demoted to grey (stale, or unverified when a venue returns no trade data). A price that stopped moving is a fossil, not a quote.
- PM 24h vol: how much was traded on the Polymarket side in the last 24 hours, a rough gauge of liquidity.
- Match: our confidence that these two rows are really the same event. A pair below 65 percent only reaches the board when Jev, the model that reads both venues' rules, signs off on it.
The board sorts four ways: by gap (biggest disagreements first), by size (most executable first), by volume, and by match confidence. The top strip shows the counts. Every row carries two link chips, PM↖ and KX↖ (with arrows), that open the exact market on each venue. Above the board, a live ticker scrolls the widest current gaps; the header chip counts down to the next scan; and when a gap moves between scans, the row flashes green or red.
Open any row (or press its arrow button) and the drawer gives the full workup: gap and venue charts (the last 24 hours, hourly, for free visitors; 90 days, daily, for desks), Kalshi volume, the settlement-terms check, and the order-book panel, with top-of-book price and size on both venues, plus the cross-venue executable figure (the smaller of the two legs, with the buy side resolved from the gap direction).
Why you can trust it
Every pair on the board passes three independent gates. First, a deterministic matcher (token overlap, sequence similarity, hard blocks on party-vs-person pairs) proposes the match. Second, Jev (a System One model) reads both markets' actual resolution rules and answers two questions: is it the same event, and do the venues settle it on the same terms? Different-event pairs are dismissed, unsure ones are held, and every pair on the board is labelled same contract, close terms or terms differ. Third, the staleness gate: either venue silent for 24 hours demotes the row.
The receipts are public. The matching audit shows every pair the matcher produces right now (the ones on the board, the ones held, and the ones dismissed) with match scores, both Jev verdicts, and last-print ages. Most terminals hide this layer. Ours is the product.
How you benefit
The insight: when two liquid venues disagree on the same event, one of them is slower to price reality. A 3¢ disagreement on an election market is information. It tells you where the crowd is wrong, where money is moving, and occasionally where a low-risk trade exists on one or both venues. Cross-venue disagreement is one of the cleanest signals in the market, because it comes from structure, not opinion.
Before gap133, that signal meant five browser tabs, two spreadsheets, and a group chat. The board replaces the pile with one screen.
The habit: verify everything on-venue. gap133 is a data terminal, not a brokerage. The venue chips exist so every number you act on can be checked at the source in seconds. Prices move; your edge depends on acting on live prices, not cached ones.
Free vs paid
Observer (free)
- Top 25 pairs by gap, 10-minute delayed
- Net (fee-adjusted), Size (executable contracts) and the Live column
- All sorting, the ticker, the stat strip
- Free gap alerts on our Telegram channel
- The matching audit, in full
- A 24-hour gap chart for each market
Founding desk ($17/month, crypto only)
- Live board, no delay: every pair on the board, not just the top 25
- Personal Telegram alerts at your own gap threshold: we message you when a pair reaches it
- 90 days of gap history per market (free visitors see the last 24 hours)
- The REST API: every pair as JSON, with your desk key
- An MCP server, so your AI agents can query the board directly
- The live Stock Tokens board
- The Chrome Ledger terminal skin
Payment: USDC or USDT (on Ethereum mainnet or Solana), ETH, SOL or BTC, sent to one of the addresses in the desk card on the terminal. The price is $17 in USD terms; for ETH, SOL or BTC we use the USD value at the time you pay. Checkout is manual: email your tx hash to hello@gap133.xyz and we send your desk key once the transaction confirms, usually within the hour. Here is exactly what happens next. The key runs for 30 days from the day we issue it, and renewing extends the same key. No cards, no auto-renewals.
Building something? The same desk key unlocks the REST API and the MCP server: matched pairs with fee-adjusted gaps, executable size, and verification gates, as JSON or as tools your AI agent can call. One key, terminal and code.
FAQ
Do I need a wallet or exchange account to use gap133?
No. gap133 is read-only. It never asks for your wallet, your keys, your seed phrase, or an exchange login. Nothing on this site can move your money. You always trade on the venues themselves, through their own apps.
Is this financial advice?
No. gap133 shows you a price difference and nothing else. What you do with it is your decision. Prediction markets involve real risk, including the risk of losing everything you stake. Nothing here is trading advice or a solicitation.
Why do prices differ between Polymarket and Kalshi at all?
Different audiences, different capital, different speeds. Kalshi is a US-regulated exchange settled in dollars; Polymarket is a crypto-native venue settled in USDC and restricted for US traders. Different participants see different news, at different times, with different risk appetites. The disagreement is structural, which is why it keeps appearing.
Is trading the gap risk-free?
No, and be suspicious of anyone who says otherwise. Capturing a gap usually means holding positions on both venues, which introduces settlement timing, fee, and venue-risk considerations. gap133 shows you the signal; managing risk stays your job.
How often does the data update?
We rescan the whole board every few minutes. Free views run 10 minutes behind. Desk keys see each scan as soon as it lands.
What does the Match percentage mean?
It is our matcher's confidence that the two rows are really the same event. Matching is done by title and meaning, and it is deliberately conservative: a pair scoring under 65 percent only reaches the board if Jev, the verification model that reads both venues rules, signs off on it, and any pair showing an implausibly wide gap (25 cents or more) is demoted automatically, because wide gaps almost always mean the questions are subtly different (a nominee market versus an election-winner market, for example). We would rather hide a real gap than show you a fake one.
What do Net and Size mean?
The gap is the edge at the touch: what you capture buying one venue at its ask and selling the other at its bid, before fees. We used to lead with the plain price difference, but when one book is wide that number means nothing (a 12c bid and 66c ask once showed a 55c difference with 1c actually tradeable), so it now sits underneath as the quote gap, for context only. When the edge is zero or negative the row says inside spread and is dimmed. Net is the edge at the touch minus both venues taker fees at the prices you would actually trade (we read the fee terms Polymarket publishes for each market, and round the Kalshi fee up per order the way Kalshi does). The net + badge means the trade survives real costs and the settlement terms match closely. Size is how many contracts you could trade on both venues at the touch: the smaller of the two order-book legs. Each contract pays $1 at settlement, and putting on both legs ties up roughly $1 per contract in total. Together they answer the only two questions that matter about a gap: is it real after costs, and is it real at size.
What does the Live column mean?
A green both live badge means both venues printed a trade within the last 24 hours. If one side has gone quiet, the row is marked stale and demoted to grey: a price that stopped moving is a fossil, not a quote. If a venue returns no trade data at all, the row says unverified and is treated the same way. Stale and unverified rows never get net + and never trigger alerts. This gate exists because showing stale prices next to live ones is how scanners manufacture fake gaps.
How do I know the pairs are really the same event?
Three gates. A deterministic matcher proposes the match and blocks questions that share words but ask different things. Then Jev (a probabilistic decision model) reads both markets resolution rules and answers two questions: is it the same event, and do the venues settle it on the same terms? Different-event pairs are dismissed, unsure ones are held, and every pair on the board is labelled same contract, close terms or terms differ. Then the staleness gate checks both venues are still printing. The full evidence trail for every pair is public on the matching audit page.
What does "terms differ" mean?
Both markets are about the same event, but the venues settle them differently in some scenarios: a different deadline, interim or acting holders counted on one venue and not the other, or an incumbent staying handled differently. Part of the gap is the price of that difference, so it is not free money, and we never mark those gaps net positive. Open the row to see both deadlines.
What is the Stock Tokens page?
Robinhood Stock Tokens trade on Uniswap pools on Robinhood Chain around the clock. The Stock Tokens page compares each pool price with the Chainlink reference Robinhood publishes for the same token, then subtracts the pool fee, because a pool can sit anywhere inside its fee without anyone being able to trade the difference. Only pools with real depth are shown; most Stock Tokens still sit in empty placeholder pools, which we count and leave out. Weekend gaps are drift against a frozen reference and usually close when trading resumes. Stock Tokens are not available to US persons and are restricted in several other places. We only show prices; we never route trades.
Why can I not see some pairs on the board?
They are in the review band: pairs the matcher proposed but the gates did not clear. The matching audit page shows all of them with their scores, Jev probabilities, and dismissal state. Transparency about what we are unsure of is a feature, not a bug.
A row disappeared or changed since I last looked. Why?
Markets close and resolve; volumes shift; the matcher improves. The board is a live view, not a history. Every market has a gap chart: free visitors see the last 24 hours, desks see 90 days.
The site looks completely different in another tab. Why?
You found the theme switcher in the header. Day is Daylight (light theme), Nite is Midnight Tape (dark theme). Chrome is the desk-keyholders skin; the padlock means it needs a paid key. Your choice is remembered on this device.
I paid. When do I get access?
Your desk key is sent after your transaction confirms on-chain, usually within the hour. Visit the unlock link in that email once and the board goes live in that browser. Each desk key runs for 30 days from when we issue it; when you renew, we extend the same key, so there is nothing to set up again.
How do I get a refund?
Email hello@gap133.xyz within 7 days of payment and we will sort it out. We would rather refund fast than have an unhappy desk holder.
Is this available in my country?
gap133 is a data service and works anywhere the site loads. But the venues themselves have their own restrictions: Polymarket restricts US traders, and Kalshi serves US residents. Check your venues terms before you trade. gap133 shows data; it does not grant trading access.
Who built this?
gap133 is an independent, bootstrapped product. No VC, no token, no roadmap promises. It makes money the honest way: people pay for the data because it is worth paying for.