I built a free duplicate-listing checker for classifieds sites. Wasn't sure it was even worth posting.

A site owner I know complains about the same three sellers every time we talk. They repost the same item under a slightly different title to reset the "newest" sort. Buyers notice. Trust drops. She'd never fix it manually because it's not urgent enough to stop and do, just annoying enough to eat at her.
I built a tool for it in a weekend because I was bored, honestly, not because I thought it'd go anywhere. Paste a listing, it checks title, price, and description against everything already posted, spits out a similarity score. That's it. No login. No dashboard.
I almost didn't share it. Felt too small and too ugly to post about.

I didn't gate it, but not for noble reasons
I have a classifieds software project on the side and part of me wanted to attach this to it. I didn't, mostly because I doubted anyone would bother signing up for something this minor. Turns out that was the right call for the wrong reason. People in this space have been burned by "moderation suites" that cost real money for something they use twice a month. Free with no account was the only version anyone was going to try.

What spread it wasn't what I expected
One admin posted in her group that she'd caught eleven duplicates in a week and finally had proof to ban someone she'd suspected for months. That's the post that moved it. Nothing I said did anything.
Then a few people I don't know started using it. And the question that came back wasn't "can you add more features," it was "does this catch user-reported ones too, or just new posts." Nobody wanted automation. They wanted evidence.

The side thing that mattered more than the tool
While building this I kept a running list of the sketchy tricks I noticed — reposting with word order swapped, prices nudged by a few dollars, cropped photos to dodge image matching. I posted the list separately, no tool attached. It's been shared more than the checker.
I don't fully know what to do with that. The tool needs me to keep it running. The list didn't need me at all once it existed. That feels like it should worry me more than it does.

Things I'm fairly sure about, not certain:
• Trust problems don't look like feature requests. Nobody puts "stop the same three people from gaming the sort order" on a roadmap, but everyone recognizes it instantly when you name it.
• A free tool in an oversold niche only works if it has zero strings. Any signup wall reads as "here's the upsell," even if it isn't one.
• What people ask to extend matters more than what they thank you for.
• Some of what you build travels further as a plain list than as software. I don't know how to feel about that yet.

Not sure if I keep building the tool out or just keep publishing the patterns instead. Has anyone else found that the non-software version of what they built outperformed the actual product? Genuinely asking, not being modest about it.