Hey Product Hunt,
Correction to my own post. The first version of this went live with a template placeholder still sitting in it, and it described the wrong product. Fixing it properly, because the mistake turns out to be on-topic.
MentionPop measures whether AI answer engines name your brand when someone asks for a recommendation in your category: ChatGPT, Gemini, Perplexity, Google AI Overviews. Not social listening, not alerts. The answer itself.
What pushed me to build it: I asked one engine the same buyer-intent question 16 times. Same wording, same day. It named a different "best tool" in most of the runs. If I had asked once and screenshotted the answer, I could have argued almost any conclusion I wanted, depending on which run I happened to catch.
We used MentionPop to run an early AI visibility baseline for DataEngPrep.tech.
Across 25 sampled observations, the brand received 0 mentions and 0 owned-domain citations, while 44 external domains appeared in the captured evidence.
That result demonstrates what we want MentionPop to provide: not just a visibility score, but evidence showing where attention went and what to improve next.
The next phase is to strengthen the site’s category content, structured answers and third-party authority—then repeat the same observation set.
This is an early sampled baseline, and we’ll continue refining the prompts and methodology as we build.
Pythagora
Super exciting to see mentionpop launch today. Congrats to the team and wishing you lots of momentum! ✨
@leon_ostrez Thanks, Leon! Really appreciate the support. I’m focused on making AI visibility evidence genuinely actionable excited to keep improving MentionPop.
@leon_ostrez Following up properly, Leon -- I owe you more than thanks.
I ran pythagora.ai through the crawler-side checks MentionPop does. Good news first: your robots.txt allows GPTBot, ClaudeBot, PerplexityBot and Google-Extended, which most sites get wrong.
The catch: fetched as GPTBot, the homepage returns 51 characters of visible text and zero links, because it renders client-side and the AI crawlers don't execute JavaScript. They reach you fine and read essentially nothing. A prerender for bot user-agents would fix it, and everything else ---schema, llms.txt, content --- sits downstream of that.
Happy to send the full output if it's useful.