Evidence engine for prediction markets
Prediction markets hand you a number and no reason. We read the reporting around the clock in 20+ languages, turn it into claims you can check, and show you the source behind every one — so you can decide what the price is actually worth.
Research only. Not financial advice. No trades are placed, and no funds are held.
The receipts
A number you cannot check is a guess wearing confidence. Ours arrives with the reporting it was built from, and with the reporting that did not survive — because what gets discarded says as much as what gets kept.
Two G7 finance ministers publicly endorsed the framework this week.
One regional survey published this week shows a small countertrend.
Ministry statement asserts the measure already passed.
Why evidence, not a score
A share pays $1 if it is right and $0 if it is wrong, so the price is the break-even probability. At 90c, being right nine times in ten only breaks even — one loss takes back nine wins. That is why a hit rate sold as a headline tells you almost nothing.
How it works
Four steps. Fully traceable. No black boxes.
Watch
Our pipeline ingests online news stories 24/7 in 20+ languages - real-time and across recent weeks - so context is never stale.
24/7 · 20+ languages · real-time + recent weeks
Extract
Each story is distilled into discrete, checkable claims, each pinned to its source, publication time, and language.
Claims, with sources attached.
Distill
For any Polymarket or Kalshi market, the evidence is weighed into a simple directional signal - YES lean, NO lean, or insufficient evidence - with a stated confidence.
A conservative research lean - not a tip.
Verify
When markets resolve, we compare our lean to the actual outcome and publish the track record. No cherry-picking.
Scored against real outcomes.
Coverage
Live coverage across every major Polymarket and Kalshi category.
Example report
Inspect the lean. Drill into every claim, then into every source behind it.
Report preview
Cross-language coverage from primary outlets converges on a YES lean: two G7 ministers have publicly endorsed the framework, market-microstructure data shows thickening depth at the implied price tier, and the only counter-evidence is a single regional outlier survey.
Claims - select one to drill down
Claim drill-down
Two G7 finance ministers publicly endorsed the framework this week.
Sources
What we will and will not claim
Our measured record lives behind sign-in, next to the sample size it was computed from. It is not a marketing number, so this page does not quote one.
We do not claim to forecast better than the price, because when we measured it we did not. What the evidence is good for is telling you when a price rests on almost nothing — a thin book, one account, or resolution wording that can settle against a reading most people would call correct.
FAQ
No. MarketClaims is a research tool. Our leans are evidence summaries, not recommendations to trade. We place no trades and hold no funds.
A discrete, checkable statement extracted from a news story - pinned to its source, publication time, and language - that bears on a market's resolution criteria.
Online news coverage monitored 24/7 in 20+ languages, spanning real-time stories and recent-weeks context. Every claim links to its source.
When evidence is thin or conflicting, we say so - the verdict can be “insufficient evidence.” We'd rather abstain than overreach.
Calls are timestamped before resolution and graded against actual outcomes - hits and misses both published.
No. There is nothing to connect. MarketClaims is read-only research - it holds no funds and connects to no wallets.
Polymarket and Kalshi markets; paste a market URL or browse by category.
We don't promise that, and you should be wary of anyone who does. We promise sourced evidence and an honest scorecard.
Pick a market. Read the evidence. Decide for yourself.