Last updated:
Open Graph Checker
This open graph checker reads a page's og: tags, checks its og:image's real size, and draws a generic link preview.
How to use the open graph checker
Enter a page's address under "Check a URL" and our server fetches it as SchemaBizBot; the HTML comes back and is checked in your browser. Under "Paste HTML" you can check markup that isn't live yet. Nothing in pasted HTML is fetched, so the image check needs a URL. "Try an example page" loads a sample with a few deliberate mistakes.
The results open on the Open Graph tab. Each finding is labeled Error, Warning, Note or OK, and the tab's badge counts the errors and warnings. Below them is a table of the page's og: tags. A URL check also shows the fetch: status, redirects, size and time, and a "Copy a link to this check" button. The Overview tab draws two generic link previews, one with a large image and one with a square thumbnail.
The required tags and the recommended ones
The Open Graph protocol (ogp.me) makes four properties required for every page. The OG tag checker reports each one that's missing as a warning.
| Tag | Status | What it's for |
|---|---|---|
og:title | Required | The headline of the preview |
og:type | Required | What the page is: website, article, book, profile, or a music.* or video.* type |
og:image | Required | The picture shown with the link |
og:url | Required | The canonical address of the page |
og:description | Recommended | A sentence or two under the title |
og:site_name | Recommended | The name of the whole site, as opposed to the page |
og:image:alt | Recommended | A text description of the image for people who can't see it |
og:locale | Optional | The language and territory, such as en_US (the default); og:locale:alternate lists other versions |
A missing og:description, og:site_name or og:image:alt is a note, and so is an og:locale not written as language_TERRITORY or an og:type that isn't an Open Graph type. On alt text the protocol is direct: a page that specifies an og:image should specify og:image:alt.
og:image: size, ratio and format
Facebook's image guidelines set the floor at 200 × 200 pixels and the ceiling at 8 MB. For the best display on high-resolution devices they ask for at least 1200 × 630, and at least 600 × 315 for the larger link previews, kept close to a 1.91:1 aspect ratio. They also ask pages to give og:image:width and og:image:height, so the image loads properly the first time the link is shared.
In the checker, an image under 200 × 200 is an error. Warnings: under 1200 × 630, SVG or any format other than JPEG, PNG, GIF or WebP, a file over 5 MB (X's limit, with Facebook's 8 MB named too), and an og:image:width or og:image:height that doesn't match the real image. A ratio more than 10% away from 1.91:1 is a note.
How the checker finds the real size
On a URL check, our server reads only the first 64 KB of the og:image, enough to get its format, width and height. The image is never passed on or kept. One that doesn't load, or isn't an image, is an error. The previews show a placeholder of the size found, not the image.
The 1 × 1 placeholder
A 1 × 1 image is a placeholder pixel, not a picture: the tag is there, but the preview has nothing to show. The checker calls this case out by name rather than reporting it as merely small.
og:url and the canonical
The protocol calls og:url the canonical URL of your object, so it should be the same address as the page's rel="canonical". When the two differ, the checker warns, and it expects absolute addresses: a relative og:url or og:image is an error. To check the canonical itself, and whether its target loads, use the canonical checker.
property= or name=, and tags set twice
Open Graph tags are <meta property="og:…" content="…"> elements in the <head>. A tag written with name= in place of property= doesn't follow the protocol, so the checker warns about it.
When a theme writes one og:title and a plugin another, the protocol settles it by order: the first tag from the top is preferred. The checker warns when a tag is set twice with different values.
A complete Open Graph block for an article
Every tag here passes the open graph validator's checks on the Open Graph tab, as long as the image really is a 1200 × 630 JPEG:
<head>
<link rel="canonical" href="https://example.com/guides/descale-a-kettle/">
<meta property="og:type" content="article">
<meta property="og:title" content="How to descale a kettle with vinegar">
<meta property="og:description" content="Four steps, one hour, and no scrubbing.">
<meta property="og:url" content="https://example.com/guides/descale-a-kettle/">
<meta property="og:site_name" content="Example Kitchen">
<meta property="og:locale" content="en_US">
<meta property="og:image" content="https://example.com/images/descale-kettle-1200x630.jpg">
<meta property="og:image:type" content="image/jpeg">
<meta property="og:image:width" content="1200">
<meta property="og:image:height" content="630">
<meta property="og:image:alt" content="A steel kettle beside a jug of white vinegar">
</head>
Both addresses are absolute and og:url matches the canonical.
Common mistakes and what the checker reports
- A relative image address.
/images/kettle.jpgin place of a full address is an error, and the example page has it. - No og:url. A missing required tag is a warning; the example page shows this too.
- og:url that disagrees with the canonical. Say, a staging domain or a dropped trailing slash; a warning.
- An image that's too small. A 150 × 150 icon is an error; a 600 × 315 image is a warning.
- Stale width and height. The image changed but
og:image:widthandog:image:heightdidn't; a warning. - Two sets of tags. A theme's and a plugin's, with different titles or images; a warning saying parsers take the first.
Where X cards and canonical tags fit
X reads its own twitter: tags first, and falls back to og:title, og:description, og:image and og:image:alt when they're missing, so complete Open Graph tags already give X a title, description and image. Its size limits and card types differ, though; the Twitter card validator checks them. The canonical checker covers rel="canonical" and robots rules. The meta tag checker runs every check at once, titles and descriptions included.
Frequently asked questions
FAQ
What is an Open Graph checker for?
For seeing what a page tells the apps that build link previews. It lists every og: tag the page has, flags the required ones that are missing or written wrongly, and, for a URL check, measures the image the preview would use.
Why does my link preview show the wrong image, or none?
The usual causes are an og:image that is missing or written as a relative address, an image smaller than 200 × 200 pixels, or two og:image tags where the first one isn't the image you meant: parsers take the first. Another cause is a page that sends a server's request different HTML from what you see in your browser. A URL check fetches the page the way a server does, so the tags it lists are the ones our server was sent.
What size should og:image be?
Facebook's documentation asks for at least 1200 × 630 pixels for the best display on high-resolution screens, an aspect ratio close to 1.91:1, and a file no larger than 8 MB. X's limit is lower: under 5 MB. Facebook's minimum is 200 × 200.
Does the checker store the pages or images it checks?
No. Pasted HTML is checked in your browser and never leaves it. For a URL check, the address goes to our server, which fetches the page as SchemaBizBot, along with its canonical target when that's another page, the first bytes of its social images and, if you check return links, up to 10 alternate versions, and sends the HTML back to your browser to check. Like every tool here, the checker records what kind of check ran, such as how many required Open Graph tags were missing and whether the image passed, never a tag's text or the page's content. A URL check also records the page's address without its query string.