How the Checker Reads Your Source
You paste the page source, which is the markup a browser receives, and the checker parses it into a document with the browser's built-in DOMParser. The parsed copy is inert. The tool does not run its scripts, and it never requests the addresses it finds in links, images or canonical tags.
Because the tool reads source text, it sees only what the paste contains. Elements added by JavaScript after the page loads are missing unless you paste the rendered markup, for example copied from the browser's element inspector.
Each check returns pass, warn or fail. A pass scores one point, a warn scores half a point, and a fail scores nothing. The points are averaged across the checks and shown as a number out of 100. That number summarizes the checklist. It is not an estimate of how a search engine will rank the page.
A pasted-source audit is a snapshot. A live crawler requests each address, follows redirects and may run scripts, but this tool sees only the text you copied at one moment. If the page changes after you copy it, run the checklist again on fresh source.
The domain field is optional. Leaving it blank does not cause an error. It means absolute links to your own site are counted as external, which can lower the internal-link result for a page that links to itself by full address.
The Checks and Their Thresholds
| Check | Rule the tool applies | Result |
|---|---|---|
| Title | Present, 30 to 60 characters | Pass in range, warn outside it, fail when empty or missing |
| Meta description | Present, 120 to 160 characters | Pass in range, warn outside it, fail when missing |
| H1 heading | Exactly one H1 | Pass with one, warn with several, fail with none |
| Heading order | First heading is an H1, and levels do not jump downward | Pass, or warn |
| Image alt text | Every img has an alt attribute; an empty alt is accepted | Pass, or warn for each missing attribute |
| Internal links | Relative links and absolute links to the domain you enter | Pass, or warn when there are none |
| Canonical link | One canonical element with a non-empty href | Warn if missing, fail if more than one |
| Language | The html element has a lang value | Pass, or fail when missing |
| Viewport | A viewport meta tag that includes width=device-width | Pass, warn, or fail when missing |
| Open Graph | Five tags: title, description, image, url and type | Pass with all five, warn with three or four, fail with fewer |
| Robots meta | Presence of a noindex directive | Pass, or warn when noindex is present |
| Word count | Visible text, with scripts and styles removed | Pass at 300 or more, warn at 100 to 299, fail under 100 |
Some checks test presence and nothing more. The canonical link and robots meta checks confirm that the element exists and has a sensible shape, but they do not judge whether the target is right, because that depends on how your site is organized. Only you can decide whether a canonical address points at the version you want indexed.
Worked Example: The Sample Page
Load the sample page and enter example.org as the domain. The sample is a short trail-shoe review with a heading that skips a level, one image with no alt attribute, a title that is short, no canonical link, and only three of the five Open Graph tags. The checklist returns:
- Title (warn): 19 characters, below the usual 30 to 60 range.
- Meta description (pass): 128 characters, inside the usual 120 to 160 range.
- H1 heading (pass): exactly one, reading Trail Running Shoes.
- Heading order (warn): one heading skips a level, because the H1 is followed directly by an H3.
- Image alt text (warn): one of two images has no alt attribute, the grip close-up.
- Internal links (pass): four internal and no external links, because the domain was entered.
- Canonical link (warn): no canonical element is present.
- Language (pass): lang is set to en.
- Viewport (pass): includes width=device-width.
- Open Graph basics (warn): three of five tags are present; og:image and og:url are missing.
- Robots meta (pass): no noindex directive.
- Word count (warn): about 109 words of visible text, which falls in the light band.
The summary reads 6 pass, 6 warn, 0 fail, out of 12 checks, and the score comes to 75 out of 100. Because the domain was entered, the four links that point to the sample site count as internal, and the fragment link is ignored entirely.
To move the sample toward a clean result, you would lengthen the title to somewhere between 30 and 60 characters, add a canonical link to the preferred address, give the grip photo an alt attribute, add the missing og:image and og:url tags, and add enough visible text to clear the 300-word band. Running the checklist again after each change shows which result moved and by how much.
Tips for Acting on the Results
- Fix failures before warnings. Failures usually break something visitors or machines depend on, such as a missing language declaration, a missing viewport or no description.
- Write titles and descriptions for people first. The length bands are guidelines, and a clear sentence that fits the page is worth more than a number.
- Use one H1 that describes the page, then step down heading levels in order.
- Write alt text that describes the purpose of each image. Decorative images can use an empty alt, which is different from a missing attribute.
- Enter your domain before running the checklist. Without it, absolute links to your own pages count as external.
- Re-run after each change. The checklist is quick, so paste the new source and compare the score.
- Do not chase a perfect 100. Some checks, such as word count, measure a rough signal rather than quality.
- Read the heading outline in order. A skipped level is a warning, and it can also show that a section needs a heading of its own.
- Pair the Open Graph tags with a real image. The og:image check only confirms that the tag exists, so confirm that the address it points to loads.
Frequently Asked Questions
Not always, but it is a good habit. A canonical link tells search engines which address you prefer when the same content can be reached at several URLs, such as with tracking parameters. The checklist warns when it is missing so you can decide.
An empty alt is a deliberate choice for decorative images, and the checker accepts it. A missing alt attribute usually means nobody made a decision, so the checker flags it for review.
A short title is not an error by itself. It is outside the usual guide, so the checker flags it for review. Only an empty or missing title is treated as a failure.
It is comparable only as a check on the same checklist. Two pages scored with this list can be compared, but the number does not say how either page performs in search.