Show IH: KinetixSEO — an SEO/GEO auditor that generates the fix and re-verifies it worked
Built this after getting annoyed that every "AI SEO" tool either scores your site or writes content for you, never both — and none of them check whether their own suggested fix actually did anything.
KinetixSEO takes a URL, scores it across classic SEO and AI-search citability (GEO — will ChatGPT/Perplexity/Gemini actually cite you), generates the specific fix for each finding, applies it, then re-scans to confirm it worked.
Ran it on our own homepage before showing anyone: 94/100 SEO, but only 41/100 on citability — the single heaviest-weighted GEO dimension. Full report's public: https://kinetixseo.com/seo-check/QQCGinF2dD
Normally €29 for a one-off full analysis (fix + re-verify). Happy to send a few of you a free code if you run it on your own site and tell me honestly what it gets wrong.
What's the most annoying "AI SEO" claim you've seen a tool make that turned out to be nonsense?
One thing I’d fix in the launch messaging: make it explicit that KinetixSEO generates the fix, but the user applies it.
In this post, the flow reads as “generates the specific fix, applies it, then re-scans.” But your current site describes the workflow differently — KinetixSEO generates the copy or code, the user applies it, then the re-scan verifies it went live.
I’d use the same three-step wording everywhere: KinetixSEO generates the fix → you apply it → KinetixSEO verifies it. That removes an important expectation gap around how hands-off the product actually is.
I also offer a lightweight $99 Launch Check for AI-built MVPs, but even if you don’t need a paid Launch Check, this is the first thing I’d fix.
I did miss that, thanks! Going forward I will use this. I'll checkout your site - which is not on you profile
Good point — I should add it to my profile. Here's the site: https://sites.google.com/view/mvplaunchcheck
Separating the audit from the fix is a strong trust model. I’d be curious whether users respond better to a single score or a field-level explanation of what changed and why. For catalog-heavy sites, title, description and structured-data issues often need different evidence, so an aggregate score can hide the highest-impact problem. How do you validate that a generated fix improved the result rather than simply changing the wording?
The entire page and content goes through multiple stages, each to ensure the latter did not forget anything or is simply mistaken. One of the steps is a critic. One that says everything is wrong based on your page and what was researched. So in short the generated fix is based on the contents of the page and research. Is it perfect? No because perfect would mean analysing the entire site, you competitors, doe keyword research and much much more. So treat it as a very good guess based upon metrics and analysis.
That makes sense. The critic and research stages sound like useful guardrails, and describing the output as a well-informed guess is more credible than promising correctness.
For catalog-heavy Shopify stores, I’m exploring a stricter version of that workflow: constrain every generated field to facts already present in the product and variant data, show the current and proposed content side by side, require merchant approval, and then rescan after the change.
Do you expose which page elements or evidence drove each recommendation? That seems especially useful when a user wants to understand why the score changed.
Yes, the full scan is over 160 checks. Each scan is saved and you can compare 2 scans. Each scan you can verify once. to check if the suggested fixen went live.
You question is one of the answers I try to give with kinetixseo.com. What is wrong, how do I fix it and how can I verify it.
The other part is currently being extended and that is AI citation and ranking
The before/after comparison is probably the strongest part of the product. A small evidence diff beside each changed score—what changed, which check passed, and what is still uncertain—would make the verification easier to trust. The AI citation and ranking extension sounds interesting too, especially if you can separate measurable page changes from the noisier external ranking effects. Thanks for clarifying.