GeminIQ

At some point I started digging into how the financial data tools I was using actually got their numbers. Almost none of them read SEC filings directly. They pull from a third-party API, someone else's interpretation of the raw XBRL data companies file with the SEC, and that layer is where things quietly go wrong. Stock splits that don't get adjusted correctly across multiple events. Inconsistent tagging that throws off a metric by a wide margin. Errors that just sit there because nobody upstream is checking.

So I built a pipeline that skips the middleman and reads the filings straight from EDGAR. My twin brother Brett came on to lead the frontend, and GeminIQ turned into a real company from there. I own the backend, the data pipeline, and the product direction; he owns everything the user actually touches. We've been building this way since day one, bouncing ideas off each other constantly, which has honestly been the best part of doing this as a team instead of solo.

The technical work has been the fun part. Chasing down a stock split bug that was corrupting balance sheet metrics. Auditing an analytics function across a dozen separate issues. Building staleness logic so the data refreshes correctly instead of silently going stale. That's the stuff I actually enjoy.

What I didn't expect: launching the product meant my job basically flipped. Before launch I was almost entirely heads-down on the pipeline. Now most of my week is marketing, content, and figuring out distribution, none of which I had any real background in going in. That's been the harder half of this by a wide margin, and probably the part every technical founder underestimates.

The product itself is straightforward. Free tier with a real set of metrics and a few years of history, paid tier for the full depth. No trying to compete with an institutional terminal on price. Just trying to be the accurate option that doesn't cost what an institutional terminal costs.

Still early. Still learning the parts of this that don't come naturally to me. But the core bet, that people want data they can actually trust the source of, is one I believe in enough to keep going.