Free AI visibility tools
Nine tools that each answer one question. Check what is broken, look up whether an AI answer cited you, generate the file that fixes it, or find out which crawler decides whether you get cited at all. Every one runs in your browser — nothing you paste is uploaded, and no email is asked for.
Check it, then fix it
AI readability check
Eight checks on one pasted page: crawler access, llms.txt, structured data, word count, FAQ block, summary block, meta description, heading structure. Red–amber–green, every result shown, including the ones that pass.
FREE Open →Fix generator
Writes llms.txt, llms-full.txt, the AI-crawler block for robots.txt, and site-wide JSON-LD in one pass, then shows you which of the eight checks each file addresses. Runs entirely in your browser.
FREE · UNLIMITED Open →Citation test
Ask a real question and see which sources the assistant relied on, and what the cited page carried that yours did not. Not open yet — a worked example from our published experiment is up now.
NOT OPEN YET Open →Generate one file
The fix generator above writes all three files at once. These three write one each, and go deeper: format validation, per-crawler switches, required properties enforced.
llms.txt generator
Builds llms.txt and llms-full.txt from a list of your pages, then checks the four things the format asks for: an H1 on the first line, a blockquote summary, a description on every link, and Optional split out. It also tells you what we measured — llms.txt correlates with citation at 0.168, under the 0.285 our sample needs before we call anything a signal.
FREE · VALIDATES Open →robots.txt generator
Six retrieval crawlers and seven training crawlers as separate switches, plus three presets. It exists because blocking GPTBot and blocking OAI-SearchBot are not the same decision — 5 of the 48 sources we measured had shut a retrieval crawler out, and they were cited far less often.
FREE · PER CRAWLER Open →JSON-LD generator
Five types with required properties enforced: Organization, WebSite, Article, FAQPage, BreadcrumbList. The validation exists because we shipped the mistake ourselves — 74 question-and-answer pairs in our JSON-LD across 30 pages, with not one of those questions visible to a reader on the page.
FREE · VALIDATES Open →Look it up
AI crawler reference
Thirteen user-agents in two tables, split by whether they decide what gets cited or only what gets trained on, with the cost of blocking each one. It ends with a checker: paste your robots.txt and it tells you which of the thirteen it allows, which it blocks, and which it says nothing about — and no rule means allowed, not blocked.
FREE · 13 CRAWLERS Open →AI citation checker
Paste a domain and it tells you whether an AI assistant cited it — looked up in the 48 pages we measured, with every run, every search rank and what the page carried. Not a scan: there is no server here, and when a domain is not in the record the page says so instead of showing a progress bar.
FREE · 48 RECORDS Open →Citation gap audit
Two failures look the same from outside: the assistant never had your page, or it read it and quoted someone else. This tells you which one you have, query by query, against the 48 sources we measured — and who took the slot where you were missing.
FREE · 5 QUERIES Open →What the free layer gives you, and where it stops
Everything a machine can tell you
- All eight checks, every result
- All four generators, no usage limit
- Every fix as copy-ready code
- Citation test, 1 per day — when it opens
- Full method and caveats published
The parts we can write for you
- Full fix list, ranked by how hard each one is
- Per-page JSON-LD for every template you have
- llms.txt and llms-full.txt, written from your content
- FAQ and summary blocks for your key pages
- Step-by-step upload instructions
$199 one-time · not open yet · you upload it yourself
Why these eight, and why in this order
Most tools in this category stop at a number. You learn that your visibility score is 34 and you close the tab, because a number does not tell you what to change. The order here is deliberate: the check tells you what is broken, the generators hand you the exact file that fixes it, and the reference tells you which of those fixes is decided by a crawler rather than by your page.
Two of the eight checks are the ones we have actually measured against citation rates. Pages carrying JSON-LD were cited 36% of the time against 10% for pages without it. Pages that allowed AI retrieval crawlers were cited 32% against 7% for pages that blocked one. But only one of those two clears the bar for being called a signal on a sample this size: JSON-LD at ρ 0.391 against a 0.285 threshold. Retrieval access sits at 0.162, under the 0.301 its own smaller sample needs. Both numbers are real; only one of them is distinguishable from noise, and we would rather tell you that than round it up.
The nine tools are also the reason this site has no signup wall. Every one of them is a browser script — there is no server behind them to bill for, so charging for them would mean inventing a reason. What we charge for is the part a script cannot do: reading your templates and writing the files for you.
Questions about these tools
Are these tools free?
Yes, and there is no usage limit on them. Seven of the eight are browser scripts: no account, no email, no card. The only metered thing on the site is citation measurement, because every run costs model calls, and that one is not open yet.
Do I have to install anything?
No. Every tool runs on the page you are already looking at. Nothing is downloaded and nothing is installed.
Do the generators need my URL?
No, and that is a limitation rather than a feature. A page cannot fetch another domain without that domain sending CORS headers, so a browser-only generator cannot read your site for you. You paste the list of pages, or the page itself, and the tool writes the file. The upside is that nothing you type leaves the tab.
Which one should I start with?
The readability check. It tells you what is actually wrong before you generate anything. Running a generator first means writing files without knowing which problem they are meant to solve.
Do you keep what I paste?
No. Every tool reads your input in the browser and never sends it anywhere. That is also why there is no upload bar to watch.
Will using all of them get me cited?
We cannot promise that. Two of the eight checks are the ones we have measured against citation rates: pages with JSON-LD were cited 36% of the time against 10% without, and pages that allowed retrieval crawlers 32% against 7%. That is a correlation on a 48-source sample, not a guarantee.