Method 2026.10.11
What is measured, and what is refused.
ZeroNoise opens the public URL you name in a browser, then fetches up to nineteen further pages on that host as HTML. It also fetches robots.txt, one sitemap, and llms.txt. The marks are arithmetic. Each work line names the URL and quotes the sentence, header, or count it used. The file lists every page it fetched, with status, title, and word count. It names no product.
What does an examination fetch?
The named URL is opened in a browser when this server has one. If the browser does not open, the HTML response is what is scored, and the memorandum says so. Up to nineteen further URLs on the same host are fetched as HTML. The first paragraph is taken from the main text, after a breadcrumb and after a byline. The thinnest of those pages that returned words is opened, and both word counts are printed. The others are not opened. Each additional page is judged for a title, a description, one H1, a word count, a list or a table, and a definitional opening. A page that fails is named. The cap is twenty pages. Rank and backlinks are not measured.
Does a lower fee change the memorandum?
There is no lower fee. Schedule 2026.10.9 charges $24.99 for every examination, in every country. The marks do not change with the fee.
Five marks
Each mark starts at 100. Deductions below are the only way it moves. The mean is the arithmetic mean of the five, rounded. A mark of 100 means no deduction in that dimension fired. It is not a rank.
- Indexability
- Can a crawler fetch the document, keep it, and find a map of the site?
- Document
- Does the HTML say, in machine-readable structure, what the page is?
- Answer engines
- Can an answer engine lift a direct response without inventing one?
- Generative engines
- Can a generative system attribute and ground a passage from this document?
- Provenance
- Does the document say who is responsible for it?
Published deductions
| Dimension | When it fires | Delta |
|---|---|---|
| indexability | Final HTTP status is not 200. | -40 |
| indexability | The final URL is not HTTPS. | -25 |
| indexability | More than three redirects occur before the final document. | -6 |
| indexability | A robots meta tag or X-Robots-Tag contains noindex. | -50 |
| indexability | robots.txt for User-agent: * disallows /. | -40 |
| indexability | An HTML 200 response has no canonical link. | -8 |
| indexability | The canonical link points at a different host. | -15 |
| indexability | No sitemap is declared in robots.txt and /sitemap.xml is not available. | -8 |
| indexability | At least one additional page on the host did not return HTTP 200. | -8 |
| document | At least one additional page on the host has no title. The page is named. Only the thinnest additional page that returned words is opened. | -6 |
| document | At least one additional page on the host has fewer than 80 words in the HTML. The page is named. Only the thinnest additional page that returned words is opened. | -6 |
| document | At least one additional page has at least 80 and fewer than 250 words in the HTML. The page is named. Only the thinnest additional page that returned words is opened. | -6 |
| document | At least one additional page has no H1. The page is named. Only the thinnest additional page that returned words is opened. | -6 |
| document | At least one additional page has no meta description. The page is named. Only the thinnest additional page that returned words is opened. | -6 |
| document | At least one additional page has an image with no alt attribute. The page is named. Only the thinnest additional page that returned words is opened. | -4 |
| document | At least one additional page has no lang attribute. The page is named. Only the thinnest additional page that returned words is opened. | -4 |
| document | The document has no title. | -25 |
| document | The title is shorter than 15 characters or longer than 70. | -8 |
| document | The document has no meta description. | -12 |
| document | The meta description is shorter than 50 characters or longer than 170. | -6 |
| document | The document has no H1. | -18 |
| document | The document has more than one H1. | -8 |
| document | A heading level is skipped (for example H1 followed by H3). | -6 |
| document | The html element has no lang attribute. | -6 |
| document | No viewport meta tag is present. | -10 |
| document | Fewer than 80 words of text are in main, or in body if main is absent. | -22 |
| document | At least 80 and fewer than 250 words of text are present. | -10 |
| document | At least one image has no alt attribute. | -6 |
| document | The final response is not an HTML document. | -30 |
| Answer engines | There is no question-shaped heading and no FAQPage structured data. | -18 |
| Answer engines | The first paragraph is not definitional (no is/are/means in its opening) and the page is not only a few words. | -10 |
| Answer engines | The document has neither a list nor a table. | -8 |
| Answer engines | At least one additional page has neither a list nor a table. The page is named. Only the thinnest additional page that returned words is opened. | -6 |
| Answer engines | At least one additional page of 80 words or more does not open with a definitional verb. The page and the paragraph are named. Only the thinnest additional page that returned words is opened. | -6 |
| Generative | llms.txt is missing. The convention is young; the deduction is moderate and published. | -12 |
| Generative | llms.txt responds but contains fewer than 40 bytes. | -8 |
| Generative | No JSON-LD structured data is present. | -15 |
| Generative | No publication or modification date is present in meta tags, time elements, or JSON-LD. | -8 |
| provenance | No author is named in a meta tag, a rel=author link, or JSON-LD. | -12 |
| provenance | No link text or href suggests an about, contact, or impressum page. | -10 |
| provenance | The HTTPS response does not send Strict-Transport-Security. | -6 |
| provenance | X-Content-Type-Options is not nosniff. | -4 |
Not measured
- No ranking, impression, or click data was requested. None is reported.
- No backlink graph, traffic estimate, or competitor set was consulted.
- Core Web Vitals field data was not read. A single fetch duration is not a field metric.
- The named URL is opened in a browser when this server has one. Up to 19 further URLs on the same host are fetched as HTML, judged, and named when a published check fails. The thinnest of those that returned words is also opened. The others are not. The cap is 20 pages.
- A mark of 100 means none of the published deductions in that dimension fired. It is not a rank and not a certificate of quality.
Fee schedule 2026.10.9
The fee is $24.99 in US dollars. Every country. Every examination. Schedule 2026.10.9. Purchasing-power parity stays withdrawn. The next examination is the same amount. There is no code. The memorandum does not change with the fee.
The country you select is recorded on the examination. It does not change the charge.
The reading
After the measurements are stored, a single model call may write a reading. It is instructed to add no facts, to recommend no product, and to say when the evidence is insufficient. Sentences that name a vendor are deleted before you see them. If the reading is unavailable, the memorandum still stands: marks, defects, and holds are computed without it.