Hey PH 👋 I'm Yash, and I have a controversial opinion: AI should not be allowed to make up numbers about your money. Wild take, I know. Two years ago I landed in Newark with two suitcases. Collected an inbox full of "unfortunately" emails (I counted; searching "unfortunately" in my Gmail returns 153 results, and yes, that's a cited claim). So last May I graduated, incorporated the same day, and built the thing I couldn't stop thinking about. Monday it's 40 feet tall in Times Square, which is objectively insane for a guy whose engineering team is AI agents and a rule. The rule: if a figure can't be traced to a primary public record, it isn't shown. Not flagged with a cute warning icon. Not shown. There was a week before launch where my own verification gate rejected 95% of everything my pipeline produced. I fixed the pipeline. The gate doesn't negotiate. (It rejected a dossier three times yesterday over one unsourced earnings date. I've never been prouder of software for refusing to work.) What you actually get: daily research briefs on your watchlist, alerts when filings and insider trades hit your names, valuation models with their assumptions in the open, and every single number linking to the SEC filing or Fed release it came from. Free tier to poke around, 7-day trial with no card, first 100 verified signups become founding members. Two things I genuinely want your take on: the citation UX; tap any number and the source filing opens; is that fast enough that you'd actually check, or does it need to be even more frictionless? And the daily brief format, would you read it before your commute, or should it be shorter? Beyond that: roast me, test it, and if you find a number in there without a receipt; comment it. Publicly. That's the whole product promise and I'll be here all 24 hours to answer for it. Trust me bro is not a source. This is.

Avarieux

It checks out.

It checks out.