Skip to main content
smnify.

SEO Audit

Enter a public page URL, or paste its HTML, to run ten on-page checks: title, meta description, H1, image alt text, canonical, Open Graph title and image, language, mobile viewport and noindex. You get a checklist score and can export the findings. It checks the page's markup only and does not predict search rankings.

Browser + server request

Free, no sign-up. In URL mode the full address you enter is sent to a fetch proxy (SMNIFY's own proxy when configured, otherwise the third-party AllOrigins or CodeTabs services) so the page can be downloaded; analysis then runs in your browser. In Paste HTML mode nothing leaves your browser.

How to use the SEO Audit

  1. Pick a tab: enter a public page URL, or choose Paste HTML and paste the full page source.
  2. Press Analyse (URL) or Analyse HTML (paste). For a URL, the page is fetched through a proxy first.
  3. Read the score and the list of checks, each marked pass, warning or fail with a short explanation.
  4. Check the extracted title, meta description and headings, then fix the items marked warning or fail in your site.
  5. Press Download findings (JSON) or Download checklist (Markdown) to keep the result or hand it to a developer.

Worked example

Sample data

Input
Paste HTML mode, a made-up bakery page:
<html lang="en">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Sample Bakery: Fresh Bread Delivered Daily</title>
<meta name="description" content="Fresh bread from a sample bakery.">
<link rel="canonical" href="https://example.com/">
<meta property="og:title" content="Sample Bakery">
<h1>Sample Bakery</h1> <h1>Fresh bread daily</h1> <h2>Our loaves</h2>
<img src="/sourdough.jpg" alt="Sourdough loaf"> <img src="/rye.jpg" alt="Rye loaf"> <img src="/shop.jpg">
Output
Score: 74
pass  Title tag: 42 characters, within 30-60
warn  Meta description: 33 characters, aim for 70-160
warn  Single H1: 2 H1 tags found, use exactly one
fail  Image alt text: 1 of 3 images (33%) missing alt text
pass  Canonical URL, og:title, HTML lang (en), Mobile viewport, Indexable
warn  og:image: missing, no social share thumbnail

The page is invented for this example; the result is what the tool returns for it. Score: title 15 + description 6 (half of 12) + H1 6 (half of 12) + alt text 0 + canonical 8 + og:title 8 + og:image 4 (half of 8) + lang 7 + viewport 12 + indexable 8 = 74 of 100. Lengthening the description, keeping one H1, adding alt text to the shop photo and adding og:image would make every check pass.

How it works

What each check looks at

Title tag: present, and 30-60 characters. Meta description: present, and 70-160 characters. Single H1: exactly one H1 heading. Image alt text: the share of images with alt text. These are common guidelines for how titles and snippets display, not rules that search engines publish.

Canonical URL: a rel=canonical link exists. og:title and og:image: Open Graph tags used for link previews on social sites and chat apps. HTML lang: the html element declares a language. Mobile viewport: a viewport meta tag exists. Indexable: the page has no robots noindex meta tag.

How the score is calculated

Each check has a weight: title 15, description 12, H1 12, mobile viewport 12, image alt 10, canonical 8, og:title 8, og:image 8, noindex 8 and lang 7. A pass earns the full weight, a warning half, a failure nothing. The score is earned weight divided by total weight, as a percentage.

score = round(100 x (pass weights + 0.5 x warning weights) / total weights)

What this score is not

The score is a checklist of basic on-page hygiene. It does not measure page speed, content quality, backlinks, search intent or competition, and a high score does not mean a page will rank. Use it to catch missing basics, then judge the page on whether it answers what people searched for.

URL mode versus paste mode

Browsers cannot read another site's HTML directly, so URL mode asks a proxy service to download the page. Pages that build their content with JavaScript may look emptier to the audit than they do to a visitor. Paste mode avoids the proxy: copy the rendered HTML (for example from your browser's developer tools) and analyse it locally.

Inputs, outputs and limits

Inputs

  • A public http or https page URL, or pasted HTML source

Outputs

  • Ten on-page checks, each pass, warning or fail
  • A 0-100 checklist score
  • Extracted title, meta description and H1/H2/H3 counts
  • Download of findings as JSON or as a Markdown checklist

Limits

  • One page at a time; the tool does not crawl a site.
  • Localhost and private network addresses are blocked in URL mode.
  • JavaScript-rendered content may be missing from what the proxy returns.
  • A proxy can time out or be rate-limited; Paste HTML always works offline.
  • Not a ranking prediction: no speed, backlink or content-quality checks.

Processing

  • In URL mode the full address you enter is sent to a fetch proxy (SMNIFY's own proxy when configured, otherwise the third-party AllOrigins or CodeTabs services) so the page can be downloaded; analysis then runs in your browser. In Paste HTML mode nothing leaves your browser.

Frequently asked questions

Where does the URL I enter go?

In URL mode it is sent to a proxy service that downloads the page for you, because browsers block direct cross-site reads. Only the URL goes to that service. To avoid it, use Paste HTML, which never contacts a server.

Will fixing everything raise my rankings?

Not on its own. These checks catch missing basics that can hurt how a page is shown or understood, but search rankings depend on many things this tool does not measure, and nobody can promise a result.

Why does the audit say my title is missing when I can see one?

The page may set its title with JavaScript after loading, and the proxy returns only the first HTML. Open the rendered page's source in developer tools, copy it, and use Paste HTML.

Why is 60 characters the title limit?

It is a rule of thumb for how much of a title search results usually show before truncating. The real cut-off depends on pixel width and device, so treat 30-60 as a guide, not a requirement.

Can I audit a page that needs a login?

No. The proxy fetches pages as an anonymous visitor. Log in, copy the page source and use Paste HTML instead, but avoid pasting pages that contain private data.

Need help implementing these improvements?

We fix the issues these checks surface: metadata, structured data, performance and security headers, shipped and verified on your stack.

Get help with your siteWeb development