First, the name
This platform has rebranded. The site now presents itself as Prediction.com — its pages, API documentation and pricing all carry the new name — while the original domain, predictionhunt.com, continues to serve the same product and still appears in places, including its enterprise contact address and some documentation links. The change looks recent and not quite finished: several headings on the site still read with the old and new names jammed together where a find-and-replace stopped halfway.
This directory keeps the entry under the name it was catalogued with, because that is the URL readers arrive on. Everything below describes the current product, which is the same product under the new brand.
The reading layer
The core consumer feature is the plainest one and the most useful: for a given event, a page that lists each outcome with the payout multiple and the best available price, attributed to the venue offering it, with a link that goes straight to that market on that venue. Where the same contract trades on several venues, you see the spread without opening three tabs.
Around that sit the extras that turn a comparison table into a habit: multi-leg combos priced across venues, a daily picks page, calculators, and a wall of venue sign-up offers with promo codes, each carrying its own terms, eligibility restrictions and risk language. The offers are affiliate placements, which is worth stating plainly because they sit in prominent positions and shape what the site puts in front of you.
The smart-money feed
The most distinctive feature is a live feed of large trades from profitable wallets. The rule for entry is explicit: a wallet qualifies once its lifetime realised profit on Polymarket clears ten thousand dollars, with a higher bar for some alerts. Each entry shows the trader's lifetime profit, the size of the fill, the side, the price, how long ago it happened, and a link to the wallet's public profile — and each event page carries a panel showing which side profitable wallets are net-buying.
The method behind it is stated on the page and is worth repeating because it defines the tool's limits. Polymarket settles on Polygon, so every fill is public: wallet, market, position, contracts and price. The platform watches those fills, looks up each wallet's all-time profit from its public profile, and admits a wallet to the feed once it clears the threshold — refreshing regularly, so a wallet that starts losing drops back out.
Two consequences follow, and the platform states both. First, this is Polymarket-derived intelligence: it tells you what the most profitable wallets on one venue are doing, not what profitable money is doing across the market. Second, it is not copy trading — no orders are placed on your behalf, and the page says so directly, alongside the sensible caution that a large position is not automatically an informed one and that prices move between a fill appearing and you acting on it.
The arbitrage scanner, and its fine print
The scanner looks for the classic paired trade: buy one side on one venue, the opposite side on another, and collect the difference because the two prices sum to less than the dollar the winning contract pays. Each opportunity shows the two legs with their venues and prices, the sum, the fee deduction, and the remaining edge — presented as cents of profit per paired contract — with the return expressed both as a percentage and annualised.
What makes this page more credible than the genre standard is the bookkeeping around it. Entries are labelled as verified or estimated, where verified means the depth was checked and estimated means it was inferred. Each window records how long both legs stayed executable, and the site publishes a distribution: on the review date almost every window had lasted more than five minutes, with a median entry window of several hours, and the longest still open after nine hundred hours. The scanner also publishes its own track record — a running figure for what could have been locked over the previous day across a stated number of windows, the capital that would have been required, and the resulting annualised return.
Then comes the sentence that matters: these are scanner observations, not user fills or realised profit and loss, and overlapping markets may share liquidity. A second note clarifies that an entry window measures how long both quotes stayed executable, not when settlement pays out. In other words the platform's own pages describe something considerably more conditional than the phrase "risk-free profit" implies — the edge is arithmetic, and capturing it depends on having capital on both venues, on the depth actually being there, and on settlement frictions that the site itself lists as a reason these gaps exist in the first place: different regulatory access, different liquidity pools, two different fee models, and one venue settling in dollars against the other in a stablecoin.
The API, which is the real product
Underneath the consumer pages sits a data service, and it is the most substantial thing here. Its pitch is the problem that has annoyed cross-venue developers for years: every prediction market speaks a different dialect. One venue uses integer cents and tickers, another decimal strings and condition identifiers, another probabilities and bids only, another American odds. Two markets that settle on the same event at the same time are described by four different titles, four identifiers and four sets of settlement terms.
The API's answer is normalisation and matching. Its documentation walks through a pipeline that ingests each venue's raw payload, parses it, compares resolution rules and close times, matches equivalent markets into a canonical group with its own identifier, normalises prices into a single probability scale, and emits one shape. From there you get a unified order book that merges several venues' depth into one ladder, matched market groups that persist when a venue relists, historical candles and trade counts on one schema across venues, bulk price lookups, arbitrage and expected-value endpoints on the paid tiers, status endpoints for freshness, and a WebSocket stream for live updates.
Authentication is a single API key header with no wallet, no signed orders and no session tokens to refresh, demo endpoints work without a key at all, and the paid plans include a server that plugs the same data into AI assistants and coding tools. The platform is explicit that this is a data and intelligence service rather than an execution one — it will tell you what the price is and where the gap is, and will not place the trade.
Why it earns a place in this directory
Most tools here answer the question "what are the odds". This one answers a broader set: what are the odds everywhere, who is moving them, and is there a gap between venues large enough to matter after fees. It is the only entry in this directory that does cross-venue matching at the data layer and publishes a live arbitrage scanner with its own limitations attached, and the paid tier structure is unusually transparent for a category where most vendors hide pricing behind a sales call.
Its own guide frames the opportunity bluntly, noting that when one large platform acquired the best-known independent aggregator earlier this year, developers who needed objective data across both major venues lost their neutral option. Whether or not that reading is self-serving, the gap it describes is real, and this is one of the few tools positioned in it.
How to use it sensibly
Read the free layers first: the odds comparison is straightforwardly useful, and the smart-money feed is genuinely interesting, provided you remember it reflects one venue's most profitable wallets rather than the whole market. Treat the arbitrage page as a research instrument rather than an income statement — the platform's own disclaimers about scanner observations, shared liquidity and execution conditions are the honest framing, not the "risk-free" label. And if you intend to build anything, start on the free API tier with the starter scripts, since the matching logic it exposes is the part that would otherwise take weeks to write. Nothing here is advice, and event contract trading can lose everything you put in.





































