4 months to go. Chrome extension live. Web search integrated. 4 users. $0 revenue. Still here.

One thing none of the answers above touch: the trust problem might not be about price or timing, it might be that a single score is a hard thing to trust when it's a verdict about someone else. I hit something similar with money data, not scam data, building Alisio. Early on I showed one confident number and people made decisions off it that didn't hold up, so now if a figure can't show what it's actually built from, the app refuses to print it rather than showing a plausible-looking guess. A TruthScore that comes with the two or three signals it weighted most, the missing footprint, the claim mismatch, whatever specifically tipped it, would probably convert better than a number alone, because people won't pay to trust a black box, they'll pay to see the reasoning and agree with it themselves.

Disclosure: I build three products, one an iOS app with model calls in production, so the API bill part is familiar.

Everyone here is pointing at the 27, and they're right. But there's a separate thing in your post nobody's picked up: right now every new user makes your position worse. Each scan costs you Gemini money, nothing converts, and you've said plainly that it's money you don't have. So distribution isn't just the missing piece - it's the piece that hurts most if it arrives before the paywall works.

Which means I'd flip the order. Get one person paying, at any price, on the traffic you already have. That's a much smaller job than finding a distribution lever, and it's the only thing that turns a new user from a cost into a gain.

Concretely, I'd keep the first scan free - that's your demo, and you need it. The web-search cross-reference is where I'd put the wall. It's the thing that took you longest, it's the part nothing else does, and it's your most expensive call. At the moment your most defensible feature is also the one you're giving away to people who have told you 27 times that free is a price they're happy with.

Eight months and still shipping, with the extension out the door, is not nothing.

Still here after 4 months with 4 users — that's more than most people who talk about building.

The hardest part isn't the code, it's knowing if anyone actually wants it before you've burned months. What was your original validation signal that told you this was worth building?

Keep going.

The biggest signal here may not be distribution yet—it’s the 27 people who used the product, gave you their email, and still didn’t pay. That creates a much more interesting conversion question than “how do I get more users?” They’ve already crossed the curiosity barrier, but apparently not the value/trust/urgency barrier. I’d be careful about treating the product as proven and distribution as the only broken piece; the free-to-paid gap may be telling you something about how the product’s value is framed or experienced. The key question is: what would have needed to happen during that first scan for paying to feel like the obvious next step?

The 27 emails are probably the highest-leverage signal here. I would not ask them generally why they did not pay. I would ask one very narrow thing: "What would have needed to happen in your first scan for this to be worth paying for today?" That forces the answer toward outcome, trust, and urgency instead of politeness.

The “distribution is broken” conclusion is the part I’d question most. 27 people gave you their email after using it, but none paid — that sounds like a useful signal about perceived value, trust, or willingness to pay before assuming you just need more users. I’d dig into that gap first.

The distribution being broken while the product works is the most honest description of early stage I've read. Same situation here — still looking for that first paying customer. The hardest part is that a working product proves nothing about whether you can reach the people who need it.