Meta tag checkers and Open Graph checkers, what each one reads and the ten second check before a page goes live

A meta tag checker fetches one page and lists the tags inside its head element, the title and description first, with the lengths counted and the missing ones flagged. An Open Graph checker does a narrower job on purpose, reading only the og tags and reporting the image, title and description a social or chat app will show when somebody shares the link. If a shared link looks wrong, the Open Graph checker is the tool. If a search result looks wrong, the meta tag checker is the tool. Running the wrong one is the usual reason people conclude their tags are fine when they are not.

Google reports 210 searches a month in the United Kingdom alone for "meta tag checker", 210 for "meta tag tester" and 210 for "open graph checker", harvested on 22 October 2026 against the United Kingdom on its own. Three near identical volumes for three tools that do two different jobs.

The four things a meta tag checker actually reports

Strip away the scoring and the traffic light colours and every one of these tools is reporting the same four groups.

• The title and the meta description. Present or missing, and the character count. Both get rewritten by Google often enough that the count is guidance rather than a rule, but a missing title, or one title repeated across forty pages, is a real finding.
• The indexing and following rules. The robots meta tag, and whether a canonical is declared and where it points. This group is the one that costs pages when it is wrong, and the full set is covered in every HTML meta tag that matters and the ones to stop using.
• The social tags. The og set and the Twitter card set, which is where these tools overlap with the Open Graph checkers.
• The leftovers. Meta keywords, meta revisit after, meta author. A checker that flags a missing meta keywords tag as a problem is telling you something about the checker rather than about your page.

Why a checker and your own browser disagree

Three mechanics account for nearly every case where a tool reports a tag you can plainly see, or misses one that is definitely there.

• The tool read the raw HTML and your browser ran the JavaScript. If a tag is injected by a script after load, a fetch based checker never sees it. Viewing the raw source rather than the inspector shows you what the fetch saw, and that is the version a crawler works from too.
• The tool followed a redirect and you did not. Checking a non secure or non canonical variant of your address can land the tool on a different page from the one you are editing.
• Something cached. Social and chat platforms keep a copy of what they read the first time. A corrected og image is very often correct in the markup and stale in the app, which is exactly why the platforms publish their own debug tools with a refresh button.

The British detail a generic checker will not flag

Two things matter more on a British site than these tools suggest, and neither shows up as an error.

The first is the language and region declaration. A site selling into the United Kingdom, Ireland and the United States with near identical English copy needs its variants declared to each other. Google's own documentation on localized versions names that exact situation, "content has small regional variations with similar content, in a single language. For example, you might have English-language content targeted to the US, GB, and Ireland", read at developers.google.com on 22 October 2026. A meta tag checker will pass a page with no such declaration at all, because it is not a meta tag fault.

The second is the og locale. Setting it to en_GB rather than leaving the default affects how a shared link reads in a British feed, and no checker scores you down for missing it. While you are in the head element, settle who the page says wrote it as well. What the documentation actually says author markup is for, and what the sameAs property does on a British site, is in person schema and author bios on a British site.

The ten second check to run before a page goes live

Five checks, in this order, catch the faults that cost something.

• Is the title unique to this page. Not good, not the right length. Unique. Duplicate titles across a template are the most common defect these tools find on a real site.
• Does the canonical point at this page. A canonical copied from a template and left pointing at the home page is the fault that quietly removes a page from search, and a checker reports it as present and correct.
• Does the robots meta tag say what you meant. A noindex left behind from staging outlives the person who added it.
• Does the og image resolve over a secure connection at an absolute address. A relative og image path is the single commonest Open Graph failure there is.
• Share the link to yourself. Paste it into whichever chat app your buyers actually use and look at the card. That catches the cache, the crop and the truncation in one go, and no checker simulates it as well as the real app does.

Which tool for which situation

Situation	The tool that fits
One page, before you publish it	A browser extension that reads the head element in place, because it checks the page you are looking at rather than a fetched copy
A whole site	A crawler that exports one row per address with title, description, canonical and robots rule, because real defects are patterns rather than single pages
A share card that looks wrong	The platform's own debugging tool, since it is the only one that can clear that platform's cache
An audit somebody else has to act on	The crawl export, not a screenshot of a scoring widget

Sort that crawl export by title and the duplicates group themselves. That one sort usually produces the first real fix list on any site over about fifty pages.

Questions about meta tag checkers and Open Graph checkers

What is the difference between a meta tag checker and an Open Graph checker

Scope. A meta tag checker reports everything in the head element, which includes the title, the description, the robots rules and the canonical. An Open Graph checker reports only the og tags and usually renders a preview of the resulting share card. Use the first for search problems and the second for share problems.

Why does my Open Graph image not show when I share the link

Four usual causes, in the order to check them. The og image address is relative rather than absolute. The image is served over a non secure connection. The file is larger than the platform accepts. Or the platform cached the page before the tag existed, in which case its own debug tool has a button to fetch the page again.

Does the meta keywords tag still do anything

Not for Google. It was retired as a ranking signal years ago and Google has said so publicly. Some checkers still list it as missing because the check is cheap to run, and a few site search products inside an organisation do read it. Outside that internal case, filling it in adds nothing.

How long should a meta title and description be

Write the title so the distinguishing words come first, with around sixty characters as a working target, because the end is what gets cut. Treat the description as the sentence you would want quoted, around 150 to 160 characters. Google rewrites both on plenty of queries, so the real job is making sure every page has its own.

Can I check meta tags for a whole site at once

Yes, with a crawler rather than a single page checker. Crawl the site, export one row per address with title, description, canonical and robots directive, then sort by title to find the duplicates.

Why does the checker show a different title from the one in my template

Usually because a plugin or theme is overwriting it, or because two components both write a title and the second one wins. View the raw source of the live page rather than the template, find which value actually shipped, then look for what produced it.

The preview none of these tools can show you

Every tool on this page shows how one surface will present your page. There is now a surface none of them reads. When somebody asks an assistant about the thing your page is about, it writes a few sentences and names a company or two, and what it reaches for is its own reading of whatever has been published, not your og image and not your sixty character title. A page can therefore pass every check here and still be missing from the answer your buyer gets. It is worth knowing what the assistants currently say about your category, with your own name left out of the question so the answer is not prompted. AI Knows Us asks those buying questions across the main assistants and records which companies get named and the sources behind each answer.