signals have cleared so far. have reached an ending. The other are still open.
The reading does not stop at the obvious listed name. It names who supplies the change, who builds on it, and who loses if it holds. Where a category is empty, the signal says so.
How a signal is scored →Horizon scanning means reading the primary record systematically, to catch change before it hardens into established fact. Health agencies have done it for decades.
Institutions do it daily, against terminals that cost $14,000–$30,000 a seat. Almost nobody else does. Not because the sources are secret — they are public and free. It is thousands of documents a day, and reading them is a full-time job.
No conviction score. No entry point. No suggestion about what to do. Every thesis names the listed companies and ETFs exposed to it, verified in that run, and then it stops.
“The edge never came from connections, and I do not trust anyone on the internet enough for that. It came from reading primary sources myself, and from noticing there were always more of them than there were hours in the day.
I built PUMPKN.AI to do that reading on a fixed rubric. It has no position and no opinion. It does not know what I think — which is the only reason I trust it.”
— Founder, PUMPKN.AI
No credit card required. Cancel anytime.
candidates assessed since launch | cleared
of the cleared signals are still open. Live theses, each with a stated catalyst and a stated way to be wrong.
Everything scored so far is there the moment you sign in. Every run since launch, every signal at every score, with its rejection log. Nothing expires because you were busy, and nothing has been pulled.
/ month
Billed annually | excl. VAT where applicable | rate locked at the price you subscribe at
No credit card required. Cancel anytime.
14-day money-back guarantee once you subscribe.
No. PUMPKN.AI publishes research. It has no knowledge of you, your portfolio or your circumstances, and every subscriber sees the same output. It names companies exposed to a development. It never suggests what to do about them.
Every claim in a signal comes from the source itself or from a source read by search during that run. A claim that cannot be grounded is left out rather than hedged. Every ticker is confirmed by search in the run it appears in, or omitted.
The full rules are here →Because the rejections are the proof, and proof should be free. Anyone can check whether a story they already know was read and disqualified, and for what stated reason. That is the part you should not have to pay to verify.
Because most of the cleared archive is inside their window today. Subscribe in a year and you still get all of them, since nothing is ever pulled from the archive — but by then their windows will have closed, and a signal read after its window is history rather than something you can still act on.
Nothing automatic. There is no card on file, so nothing renews and nothing is charged. If you want to keep going you subscribe, and the rate you subscribe at is the rate you keep. There is a 14-day money-back guarantee from that point.
One person, who built it to do reading he could not do himself. The rules the engine runs on, the sources it reads, and the ways it is forbidden to fail are published in full.
Read the method →No credit card required. 14 days free. Cancel anytime.