SEO Compare
Fetch two public URLs and compare title, description, canonical, robots, Open Graph and Twitter tags field by field.
What the comparison reports
Two public pages, one row per head field: title, meta description, canonical link, robots directive, page language, the five Open Graph tags a card needs and the four Twitter card tags. Each row shows both values and a verdict — same, differs, or missing on one side — so the answer is a list of what to fix, not two walls of metadata to read side by side.
It reads the raw HTTP response, exactly as a crawler that does not run JavaScript sees it. A client-rendered page can therefore look empty here, and that is a true finding about the server response rather than a bug in the tool.
How to use it
Paste two URLs and press Run. Redirects are followed (up to five, each one re-checked) and the final URL is reported for both pages, so a comparison is never made against a page you did not mean.
Safety
This is a server fetching URLs you supply, so it is guarded: every host is resolved and refused if it points at a private, loopback or reserved address, on the entry URL and on every redirect hop. Credentials in a URL are refused, the read is capped at 512 KB per page and neither body is stored.
Limits
Two URLs per run, HTML only — a PDF or an image answer is refused with its content type, because it has no head to compare. Values are compared exactly as written; a canonical that differs only by a trailing slash is reported as a difference.
Paste two public URLs and the page fetches each one's raw HTML, parses its head, and compares fourteen fields: title, description, canonical, robots and language, five Open Graph tags and four Twitter tags.
Each field gets a verdict — same, differs, missing on one side, or missing on both — and the response also reports each page's status code, final URL after redirects, whether it redirected, and whether the read was truncated at the size cap.
It reads what a crawler reads from the server response. JavaScript is not executed, so a page that renders its head in the browser looks empty here, and that is a true finding about the HTML that was served rather than a failure of the comparison.
What is fetched and how
Both requests go through the site's guarded HTTP client, which refuses a host resolving to a private address and checks every redirect hop before following it. The chain is capped, the connect timeout is five seconds and the whole request fifteen, and each body is read up to 512 KB.
A non-HTML response is refused by name, because a PDF or an image has no head to compare. Nothing is written to disk and neither body is logged; the page receives the parsed fields, not the documents.
- 14 fields: 5 core, 5 Open Graph, 4 Twitter
- Raw HTTP response, no JavaScript execution
- 512 KB per page, guarded redirects
What a difference means
Values are compared after the parser collapses whitespace, and an empty value is treated as absent. This tool does not normalise URLs, so a canonical that differs by a trailing slash is reported as a real difference, which is exactly the thing an editor is looking for.
A missing field on one side is not automatically a defect: a page may legitimately omit a Twitter card. The verdict column tells you which side is silent so you can decide whether that silence is intended.
How to use the comparison form
Paste the two public URLs into the first and second fields and press Run; each page is fetched and its head parsed on the tool server. The result is a fourteen-row table with one verdict per field: same, differs, missing on one side, or missing on both.
Read the status codes and final URLs above the table before trusting a verdict, since a redirect or a 401 explains half of all surprises. A row flagged as truncated means the head exceeded the 512 KB cap and later fields were never seen.
What to Expect
Best for
- Checking that two language versions carry matching metadata
- Finding a canonical or robots tag that differs between environments
- Auditing Open Graph and Twitter fields before a launch
Not ideal for
- A client-rendered page whose head is written by JavaScript
- Comparing rendered content, which this tool never sees
- A page behind a login or on a private address, which the guard refuses
What this tool keeps
- The final URL and status code of each page
- The raw field values, before any normalisation of URLs
- An explicit verdict for every one of the fourteen fields
What may need cleanup
- A blank comparison caused by a page that builds its head in the browser
- A trailing-slash canonical difference, which is real but may be intended
- A truncated read, which is reported and can hide fields late in a large head
Common errors
- Comparing a production page with a staging page that returns 401
- Expecting the tool to run JavaScript and find a client-side title
- Reading a missing-on-one-side verdict as a broken tag on both pages
Example use cases
- A pre-launch metadata check across two environments
- Comparing an English and a translated page
- Finding which of two templates lost a canonical
Sample input
The two URLs https://example.com/ and https://example.com/fr/ pasted into the first and second fields.
Sample output
A field table with a verdict per row, both status codes and final URLs, and a headline naming the fields that differ.
Who this is for
- SEO specialists auditing a migration
- Developers comparing two environments
- Editors checking translated metadata
Frequently asked questions
Which fields does the SEO comparison check?
Fourteen: title, description, canonical, robots and language, five Open Graph tags and four Twitter tags, compared between two public URLs.
Does it compare rendered content?
No. It fetches each page's raw HTML, so a head written by JavaScript is invisible.
What does a difference mean?
It is a real difference in the raw field values, which may still be intended, such as a trailing-slash canonical.
Build custom tools for your store
Join the SDK program to create useful ecommerce utilities for your store.