Paste a page and get a 0–100 score for whether AI answer engines (ChatGPT, Perplexity, Google AI Overviews) will cite it — measured against real signals: answer-first structure, question headings, stats, tables, FAQs, freshness. You get specific fixes, sorted by impact. The 11 scoring weights are published on the page, so it's inspectable, not a black box. Runs entirely in your browser: nothing uploaded, nothing stored, no signup. Free, and staying free. Built by an SEO of 11 years.
No reviews yetBe the first to leave a review for Citability Scorer
Maker
📌
Hi Product Hunt 👋 I'm Marco. I've done SEO for eleven years, mostly on my own content sites, and lately I've been trying to figure out GEO — getting accurately cited by AI answer engines — in the open.
One thing kept happening: I'd look at a page that should get cited and it wouldn't, and the reasons were almost always structural — the answer was buried, the headings didn't match how people ask, there were no concrete numbers to quote. So I turned the checklist I was using by hand into a tool.
You paste a page (text or Markdown), it scores it 0–100 against eleven weighted signals and hands you the fixes sorted by impact. It runs entirely in your browser — your content never leaves your device, nothing is stored, no signup.
What it can't do: promise you a citation. Nothing on-page can — authority and third-party mentions matter too, and no tool can see those. It measures the structural things you actually control. I published exactly how the score is calculated, with the weights, because a scorer you can't inspect isn't worth much.
It's free and it'll stay free. There are three more tools in the same set (an AI-crawler checker, an llms.txt generator, an FAQ-schema generator). Happy to answer anything — and genuinely interested in where the heuristics feel wrong to you.
Report
Would love a competitor URL diff mode where I can paste two pages and see which one scores higher and exactly which signals flipped between them. Makes it way easier to justify rewrites to clients.
Report
Maker
@pekkoalifapm This is a great call, and honestly the most useful piece of feedback I could have gotten today.
A diff mode makes total sense the "which signals flipped" part especially, because that's the version of the output you can actually put in front of a client. A number is an opinion; "your competitor answers the question in the first sentence and you bury it in paragraph three" is an argument.
The tricky bit is fetching a competitor's page: right now everything runs locally in your browser and nothing is uploaded, which I want to keep. So it'd likely mean pasting both pages rather than URLs, at least in v1 less magic, but it keeps the privacy promise intact.
I'm going to build it. If you want to be the one who tells me it's wrong before anyone else sees it, drop me a line at marco@marcodiversi.com and I'll send it over when it's ready.
There's now a "Compare two pages" tab on the same page. Paste your page and theirs, and you get both scores plus a per-signal diff all eleven checks, side by side, sorted by the biggest gap first, so you can see exactly which signals flipped. There's a "Copy comparison" button that spits out clean plain text you can drop straight into a client email.
I kept it paste-based rather than URL-based on purpose: everything still runs locally in your browser and nothing gets uploaded, and I didn't want to trade that away for the convenience of fetching a URL.
Your framing is what sold me on it a number is an opinion, but "they answer the question in the first sentence and you bury it in paragraph three" is an argument. That's the version a client actually acts on.
Tell me where it's wrong.
Report
Really appreciate that the weights are public and it runs locally, finally a tool I can actually trust. Tried it on a few posts and the suggestions lined up with what I'd want an AI engine to pull from.
Report
Maker
@kuzey65xo Thank you trust was the whole reason I published the weights. I've used enough SEO tools that hand you a number and expect you to take it on faith, and I never liked being on the receiving end of that.
Good to hear the suggestions matched your instinct on what an engine would pull. That's about the best signal I could ask for, since the heuristics are essentially my own read of what gets quoted, written down and made checkable. If you ever hit a page where it disagrees with you and you think it's the tool that's wrong, I'd genuinely like to see it.
Report
The inspectable scoring weights are genuinely refreshing, honestly most tools like this just hand you a number and hope you trust it. Ran a few of my own posts through it and the fix suggestions were kind of annoyingly accurate, especially the nudge to add real stats instead of vague claims.
Report
Maker
@sla919522062231 "Annoyingly accurate" is going straight on the landing page 😄
The stats nudge is the one I fight with on my own writing the most. It's easy to write "AI search is growing fast" and much harder to go find the actual number but the number is the thing an engine can lift and attribute, and the vague claim isn't. The tool basically exists to make me stop taking the easy option.
Thanks for running your own posts through it that's exactly the use I hoped for.
Report
Pasted a blog post I'd been tinkering with and the score felt honestly harsh but accurate. The "specific fixes, sorted by impact" framing is what sold me — most tools dump a vague checklist and call it a day.
Report
Maker
@bedriyedoa3hpc "Harsh but accurate" is the goal thank you. I'd rather the tool tell you the uncomfortable thing now than let an AI engine quietly skip your page later. And yeah, the sorted-by-impact part is the whole point: a flat checklist makes you guess what matters, and usually the thing that matters most is the boring one you were avoiding. Appreciate you running a real post through it.
Would love a competitor URL diff mode where I can paste two pages and see which one scores higher and exactly which signals flipped between them. Makes it way easier to justify rewrites to clients.
@pekkoalifapm This is a great call, and honestly the most useful piece of feedback I could have gotten today.
A diff mode makes total sense the "which signals flipped" part especially, because that's the version of the output you can actually put in front of a client. A number is an opinion; "your competitor answers the question in the first sentence and you bury it in paragraph three" is an argument.
The tricky bit is fetching a competitor's page: right now everything runs locally in your browser and nothing is uploaded, which I want to keep. So it'd likely mean pasting both pages rather than URLs, at least in v1 less magic, but it keeps the privacy promise intact.
I'm going to build it. If you want to be the one who tells me it's wrong before anyone else sees it, drop me a line at marco@marcodiversi.com and I'll send it over when it's ready.
@pekkoalifapm Update: it's live.
There's now a "Compare two pages" tab on the same page. Paste your page and theirs, and you get both scores plus a per-signal diff all eleven checks, side by side, sorted by the biggest gap first, so you can see exactly which signals flipped. There's a "Copy comparison" button that spits out clean plain text you can drop straight into a client email.
I kept it paste-based rather than URL-based on purpose: everything still runs locally in your browser and nothing gets uploaded, and I didn't want to trade that away for the convenience of fetching a URL.
Your framing is what sold me on it a number is an opinion, but "they answer the question in the first sentence and you bury it in paragraph three" is an argument. That's the version a client actually acts on.
Tell me where it's wrong.
Really appreciate that the weights are public and it runs locally, finally a tool I can actually trust. Tried it on a few posts and the suggestions lined up with what I'd want an AI engine to pull from.
@kuzey65xo Thank you trust was the whole reason I published the weights. I've used enough SEO tools that hand you a number and expect you to take it on faith, and I never liked being on the receiving end of that.
Good to hear the suggestions matched your instinct on what an engine would pull. That's about the best signal I could ask for, since the heuristics are essentially my own read of what gets quoted, written down and made checkable. If you ever hit a page where it disagrees with you and you think it's the tool that's wrong, I'd genuinely like to see it.
The inspectable scoring weights are genuinely refreshing, honestly most tools like this just hand you a number and hope you trust it. Ran a few of my own posts through it and the fix suggestions were kind of annoyingly accurate, especially the nudge to add real stats instead of vague claims.
@sla919522062231 "Annoyingly accurate" is going straight on the landing page 😄
The stats nudge is the one I fight with on my own writing the most. It's easy to write "AI search is growing fast" and much harder to go find the actual number but the number is the thing an engine can lift and attribute, and the vague claim isn't. The tool basically exists to make me stop taking the easy option.
Thanks for running your own posts through it that's exactly the use I hoped for.
Pasted a blog post I'd been tinkering with and the score felt honestly harsh but accurate. The "specific fixes, sorted by impact" framing is what sold me — most tools dump a vague checklist and call it a day.
@bedriyedoa3hpc "Harsh but accurate" is the goal thank you. I'd rather the tool tell you the uncomfortable thing now than let an AI engine quietly skip your page later. And yeah, the sorted-by-impact part is the whole point: a flat checklist makes you guess what matters, and usually the thing that matters most is the boring one you were avoiding. Appreciate you running a real post through it.