Every AI research assistant cites sources. None keeps a record of what its answers said and whether it turned out right. QL Intelligence does. Each finished answer is stamped — the price and engine signal of every ticker it names pinned at that moment, its numeric claims graded against the live data it pulled, the whole record SHA-256 hashed — and every bullish or bearish stance it takes becomes a forward-only call graded by the tape. Hits and misses, equal billing.
Answers their owners chose to publish — question, pins, grade and calls, with then → now on the permalink. Private answers never appear here.
When an answer finishes, the live price and engine signal of every ticker it names are pinned at that moment, the answer's numeric-claim grounding grade is attached, and the record is SHA-256 hashed over its canonical fields. The permalink shows the hash so anyone who saved it can detect a later edit.
When the answer takes a side on a pinned ticker — a sentence that names the ticker and argues bullish or bearish without hedging — that stance becomes a forward-only call entered at the pinned price. It settles outright at a ±10% signed move or at T+30 days with a ±2% push band, and the verdict enters the public QL Record hash chain.
No. The scoreboard exposes only ticker, direction, dates, entry price and verdict. Question text and the answer itself stay private unless you publish that answer, which creates a public permalink.
Every permalink shows then → now for each pinned ticker, and if you re-ask the same question from the permalink the new answer is linked to the old one and shows which figures changed, were dropped, or are new.
Because a record that only shows hits is marketing. Misses are listed as prominently as hits, the hit rate excludes pushes and stays null until something has graded, and nothing is backfilled or rewritten.
Not investment advice. The record exists so the assistant's claims can be checked against what happened next — the one thing no citation can provide.