Paste the articleFrom Word, a blog editor or a published page; any of them works.

The fields fill themselvesFocus phrase, slug, title tag and meta description are taken from the article.

Close the gapsEach gap says what to do. The score updates as you fix them.

Audit a published article

paste the address and we fetch the page

The page is fetched through our server; only the address is sent, never the text you type. On sites that are built in the browser, such as Wix or Squarespace, the text may not come through; in that case paste the article into the box below or upload it as a file. If you are auditing an unpublished draft, leave this field empty and paste the text into the box below. If you are checking a change you just saved, wait a few minutes: because of site and proxy caches, an older version may be served for a while. The checker warns you when the fetched page is identical to the previous one.
Which pages? The checker is written for medical content pages: articles about a disease, a symptom, a treatment or a patient question. Home, about, contact and pricing pages follow different rules; auditing them gives misleading results.

1Your article

empty

or drag the file here. Word (.docx), plain text (.txt), markdown (.md) and HTML are accepted. The file is opened in your browser and sent nowhere. PDF files cannot be audited; open the PDF and copy the text, or save it as a Word file.

You can paste plain text, markdown (# heading) or HTML. If you paste the full source of a published page, the title tag, meta description and address fill in automatically. No data is sent to a server; the audit runs entirely in your browser.

2Fields

waiting

Where do you practice?

fields you leave empty are not audited

Focus phrase Slug (end of the URL) Title tag Meta description

3Result

waiting

–

Waiting for an audit

Paste an article into the box on the left.

Passing 0 Missing 0 Critical 0

How to read the score

The number alone is not enough. Some of the 25 rules count as critical. If one of them is missing, a high score does not make up for it, just like a single critical value on an otherwise normal lab panel.

22-25, no critical gap	Ready to publish
18-21, no critical gap	Publishable, close the gaps
13-17, no critical gap	Needs work
1 critical gap	Fix it first, whatever the score
2 or more critical gaps	Do not publish as is

The numbers in the table are for 25 rules. When a rule is not measured, the denominator shrinks and the thresholds drop in the same proportion: 14 out of 19 falls in the same band as 18 out of 25.

What does "not measured" mean? If the focus phrase field is left empty, the six rules that depend on it (H1, opening, subheadings, body repetition, slug and meta description) cannot be measured. They are not counted as missing and do not lower the score; they drop out of the denominator, so the result shows out of 19. As soon as you enter the focus phrase, all six are measured and the denominator returns to 25.

25/25 does not guarantee a good article; it shows that every audited rule is met. The checker does not predict rankings.

Evidence and sources

–

Link and text hygiene

–

Advertising rules scan

–

Full version

lite version

The lite version tells you what is missing. The full version also shows how to fix each gap, runs the advertising rules scan, the schema link map, the evidence and hygiene panels and the detailed report. The audit and the summary are free in both versions. The access code is printed on the last page of the handbook.

Access code

The code is stored in this browser, so you do not have to enter it each time. If the code does not work, write to info@uzclinic.com.

* The page is read as it appears in your browser, so the result is correct on every platform. With this method, too, the text is not sent to a server; it goes from the clipboard straight to the checker.

Passage analysis

–

A model quotes a passage, not a page. The article was split at its subheadings and each section was judged on its own: does the passage tie to a question, does it have support, can the section be understood alone, is its length quotable.

Not every heading has to be a question. A passage can tie to an intent in two ways: either the heading is a question, or the first sentence directly answers what the heading names. One is enough. Structural sections such as references, contact and the byline are not scored and do not count toward the average or the below-70 count. Section headings that have subheadings of their own are also left out of these criteria.

Support here means citing sources and guidelines. Some tools suggest "put a number in every passage" or "say we measured it"; in a physician's article that path leads to made-up figures and patient testimonials, and both are wrong.

Quotable questions

–

Questions this article already answers. On AI surfaces you can be chosen as a source for these questions.

If you want to measure it: ask ChatGPT and Gemini these questions yourself and note whether your name appears. The checker finds the questions; the answers are yours to check.

Schema link map

–

The page's structured data (JSON-LD) is drawn here as nodes and arrows. This is how search engines and AI surfaces read the page: which person the article belongs to, which organization that person belongs to, which address the organization is at. What slips by in the code shows up here: entities that are not linked to each other, references to an @id that is not defined on the page, the same person defined twice in two separate records. This section does not count toward the 25-rule score; it is a separate view.

The map reads only the JSON-LD blocks on the page; Microdata and RDFa markup is not counted. Click a node to open its record below. As a rule: the author is a single Person record and every article links to it by @id; the organization is a single Organization or MedicalClinic record, and the person links to it.

Changelog Recent changes to the checker

November 1, 2026 The crawler access box recognizes the shared AI training preference line in robots.txt (__usage_ai-training) and explains its possible effect on Gemini visibility.
October 8, 2026 In the passage analysis, references, contact and byline sections are no longer scored and do not count toward the average or the below-70 count. An FAQ block now ends at the first structural heading.
October 7, 2026 The crawler access box now says that if Googlebot is blocked, Brave Search will not crawl the page either.
October 1, 2026 Wix tables merged cell text and produced a false "words run together" warning; fixed. The footer note is worded more clearly.
September 30, 2026 The score is always shown out of 25, with the number of measured rules below it. A language check marks rules that cannot be measured in a non-English article as "not measured". FAQ answers are no longer checked for a cited source.
September 29, 2026 The checker went live, together with the link check and the AI visibility test.

The 25 rules in the handbook did not change; these changes concern how accurately they are measured.

Medical Visibility Checker