SEO Checker screenshot

SEO Checker

Audit a public page's static HTML in one scored report: metadata, headings, links, structured data and visible text, 0-100. Raw response only, no JavaScript.

Privacy First
0 Runs
Updated 2026-11-07
Runs on the tool server
No account needed
Private addresses refused
Nothing stored from the response

One audit instead of five checks

A page's search-facing shape lives in five places: the head metadata, the headings, the links, the structured data and the visible text. This page reads one public URL and scores all five in a single report, each section with a pass, warn or fail status and the findings behind it, plus a 0–100 score weighted across the sections.

How to use it

Enter a public https:// URL and press Audit the page. You get the status code and any redirects that were followed, the five scored sections, the overall score, and the scope block that states what was actually read.

It scores the fetched static HTML

The checker reads the raw server response, the same bytes a crawler without a JavaScript engine receives. It does not consult Google's index, it does not run the page's scripts, and the score is not a ranking prediction. A single-page application ships an empty shell and writes its words from script, so it will score thin here; the report flags that shape as possibly client-rendered instead of calling the page bad.

What each section checks

Metadata measures the title and description against the character budgets this site itself enforces, and names a missing or duplicated canonical, a missing robots directive and a missing page language. Headings wants exactly one h1, ordered levels and no empty headings. Links counts same-origin against external and refuses empty or unresolvable addresses. Schema wants JSON-LD blocks that parse. Text wants substantial visible words and alt text on images.

A real capture

This is the recorded output for https://208.67.222.222/, a host that answers 404 with a title and twelve visible words — the thin-but-honest shape this tool exists to name: metadata 65, headings 100, links 100, schema 95, text 90, score 87.

Privacy and safety

The server fetches the URL you name, so this page does not claim to run in your browser. Private, loopback and reserved addresses are refused before a connection is opened, and every redirect hop is checked again. Redirects are capped, the connection and the whole request each have a timeout, and at most 512 KB of the response is read. The response is parsed in memory and is never written to disk or logged.

Enter a public URL and this page fetches its raw server response and scores it in five sections — metadata, headings, links, structured data and visible text — each pass, warn or fail with the findings behind it, plus a 0–100 score weighted across the sections.

It scores the fetched static HTML: not Google's index, not the JavaScript-rendered DOM, and never a ranking prediction. A single-page application served as an empty shell therefore scores thin, and the report flags that shape as possibly client-rendered rather than calling the page bad.

The request is made by the tool server. Private and reserved network addresses are refused at the first URL and at every redirect hop, the chain is capped, both timeouts are set, and at most 512 KB of the response is read.

What each section checks

Metadata measures the title and description against the character budgets this site itself enforces, and names a missing or duplicated canonical, a missing robots directive and a missing page language. Headings wants exactly one h1, ordered levels and no empty headings.

Links counts same-origin against external and refuses empty or unresolvable addresses. Schema wants JSON-LD blocks that parse. Text wants substantial visible words and alt text on images.

  • Metadata 30, headings 20, links 20, schema 10, text 20
  • Section status: pass at 90, warn at 60–89, fail below 60
  • Duplicate head tags listed once, not once per section

How the fetch is bounded

Redirects are followed by hand so each hop can be validated, and the chain stops at the shared redirect cap. Connect and total timeouts are both set, and the response read is capped, so a slow or huge page cannot hold the request open indefinitely.

A refusal, a transport failure or a non-HTML response produces a message rather than a fabricated report. Nothing from the response is written to disk or logged.

How to use the URL field

Enter the public URL and run the audit. The tool server fetches the raw response with no JavaScript running, reading at most 512 KB through the hop-validating reader; private addresses are refused at the first URL and every redirect.

Read the five scored sections, metadata, headings, links, structured data and visible text, each pass at 90, warn at 60 to 89, fail below, weighted into the 0-100 score. A thin score on a script-rendered shell is a true finding about static HTML, not a tool failure.

What to Expect

Best for

  • Scoring a page's static HTML before publishing
  • Finding a missing description or a duplicated canonical
  • Confirming headings, schema and alt text in one pass

Not ideal for

  • Checking a client-rendered page's words, which require JavaScript
  • Anything behind a login or on a private network address
  • Predicting a ranking or reading a search index

What this tool keeps

  • The values the raw response contains, with missing ones named
  • The final URL after the redirect chain
  • A scope note stating that no JavaScript was executed

What may need cleanup

  • A thin score that belongs to a client-rendered shell
  • A title or description over this site's own budgets
  • A redirect to a URL you did not expect

Common errors

  • Reading a thin report as a tool failure rather than a true finding
  • Expecting the checker to execute the page's scripts
  • Testing a localhost or private address, which is refused

Example use cases

  • Pre-publish audit of a new article or landing page
  • Finding why a page looks bare to a non-executing crawler
  • Checking title and description lengths against this site's budgets

Sample input

Enter the public URL of a product page and run the audit.

Sample output

Five scored sections with named findings, a 0-100 score, the redirect chain and a scope note about JavaScript.

Who this is for

  • SEO and content teams auditing before publish
  • Developers checking a page's static shape
  • Anyone whose page looks wrong to a crawler

Frequently asked questions

What does the SEO checker score?

Five sections of the fetched static HTML — metadata, headings, links, structured data and visible text — each pass, warn or fail, with a 0-100 score across them.

Does it read what Google sees or what JavaScript renders?

Neither. It scores the raw server response without running JavaScript, so a client-rendered page looks thin here, and the score is never a ranking prediction.

Where do the title and description budgets come from?

From this site's own SEO rules, supplied with each run, so the report cannot drift from the limits the site enforces.

Developer program

Build custom tools for your store

Join the SDK program to create useful ecommerce utilities for your store.