Link Preview Image Size: 1200×630 for LinkedIn, Facebook and X

Published 17 October 2026 · by Tekiba

The image behind a shared link

When someone pastes a link into a post, the site they post on fetches the page and looks for an og:image tag in its head: the Open Graph image, usually shortened to OG image. That image becomes the large picture above the title in the preview. One tag serves all three services: X reads its own twitter:image tag first and falls back to og:image.

We read each service's page for site owners on the publication date above. They agree on the shape far more than on the numbers.

Facebook

Meta's Images in Link Shares page, written for webmasters, says:

Use images that are at least 1200 x 630 pixels for the best display on high resolution devices. At the minimum, you should use images that are 600 x 315 pixels to display link page posts with larger images.

Try to keep your images as close to 1.91:1 aspect ratio as possible to display the full image in Feed without any cropping.

The same page sets a file-size ceiling of 8 MB, and explains that images are cached by their address: change the picture but keep the file name, and Facebook can go on showing the old one. It also recommends the og:image:width and og:image:height tags, so the first share can show the image before the crawler has finished with it.

LinkedIn

LinkedIn's help article Make your website shareable on LinkedIn lists the same tags and gives, for the image:

Minimum image dimensions: 1200 (w) x 627 (h) pixels

Recommended ratio: 1.91:1

with a maximum file size of 5 MB. 1200×630 is three pixels taller than LinkedIn's minimum and the same shape to within a rounding error, so one image does for both. Further down, the page adds:

Images less than 401 pixels wide display as a thumbnail image.

X

X is the awkward one. The developer pages that used to describe its cards now redirect to the front page of its documentation, so there is no current page to quote. The last published version of its Summary Card with Large Image page, as kept by the Internet Archive, said:

Images for this Card support an aspect ratio of 2:1 with minimum dimensions of 300x157 or maximum of 4096x4096 pixels. Images must be less than 5MB in size.

2:1 is a little wider than 1.91 to 1. A 1200×630 image shown in a 2:1 frame loses a thin strip, and nothing on the page said whether from the top, the bottom or both. Treat that as one more reason to keep the edges clear rather than as a reason to make a separate image for X.

Where a crop falls, and the safe area

None of the three pages says which part of an image survives a crop. What can be said is where the risk is:

• Every edge. A 2:1 frame is wider than a 1.91 to 1 picture, so something has to give at the top or bottom; an image taller than 1.91 to 1 has more to lose; and an image under LinkedIn's 401-pixel width becomes a thumbnail, where lettering is too small to read anyway.
• The corners most of all. A corner is on two edges at once, and it is where a watermark usually goes.

As a precaution, then, keep the title and anything else that has to be read in the middle of the image, with a clear margin all round, roughly a tenth of the width at each side and a little more at the top and bottom; that is our rule of thumb, not a figure from any of the three pages. Set the text large: a preview is shown at a fraction of 1200 wide on a phone, and thin type is the first thing to go.

A watermark on a link preview

A link preview usually carries your site's name beside the image already, so a mark on it is less about credit and more about the image being reused elsewhere. If you add one, put it inside the safe area: bottom right but pulled well in from the corner, or the middle of the bottom edge. Keep it small; the preview is there to make someone click, and a mark across the picture works against that. As everywhere, a watermark only discourages copying; anyone can crop or paint over one.

Making a batch of 1200-wide images in your browser

SignetDrop resizes, pads and marks; it does not crop, and 1.91 to 1 is not one of its shapes. For a folder of article photos:

1. Size to 1200 wide. On the free version, set Resize to Width and the amount to 1200. A 4032×2268 phone photo in the wide 16:9 mode comes out at 1200×675, already close to the preview shape. A 4032×3024 photo in the usual 4:3 comes out at 1200×900, which is too tall for a 1.91 to 1 frame: some of its height has to go, and none of the pages says from where. Keeping the subject away from the top and bottom is a precaution, not a promise; check the preview each service offers before you rely on it.

2. Or pad to a wide shape. Setting a shape is a paid setting, and it adds a plain border instead of cropping. A 4032×3024 photo sized to 900px on its long edge and then set to 16:9 comes out as 1200×675, with the whole photo in the middle and bands of your background colour at the sides.

3. Place the mark inside the safe area. Choose the bottom-right or bottom-centre anchor and raise the margin until the mark sits well clear of the corner.

4. Name the files for their address. Because Facebook caches an image by its address, a changed image needs a new name. On the free version, Original name plus a suffix with -og gives hero-og.jpg from hero.jpg; change the suffix when you change the picture.

The photos never leave your machine: they are drawn onto a canvas in the page, saved again and put into a ZIP built in the tab. By default the copies carry no EXIF, so no GPS location goes out with your link previews. JPEG at the default quality keeps a 1200-wide photo well under every service's file-size limit.

What this does not do

It prepares the image file. It does not write the og:image tag into your page, crop to 1.91 to 1, or set text on the image, and how each service crops and caches a preview is its own decision, not ours; their pages can change, and X currently has none. Set shapes and output above 2048px are in the paid version, $29 once, with the full split on the pricing panel, and SignetDrop never writes its own name onto your photos.

Questions