Make the image that shows when someone shares your link. Type a title and a subtitle, pick a theme, drop in a logo or a background photo, and export a PNG at exactly 1200×630 — the size every platform crops its preview from. The title resizes itself to fit whatever you type, so a long headline shrinks and wraps instead of running off the edge, and the contrast between your text and background is checked as you go, because a share card is read at thumbnail size in a crowded feed. Everything is drawn in your browser: your logo and your background photo are never uploaded.
| Platform | Size | Note |
|---|---|---|
| Open Graph / X | 1200 × 630 | The universal default |
| 1200 × 627 | Practically the same crop | |
| Instagram feed | 1080 × 1080 | Square |
| Instagram portrait | 1080 × 1350 | Tallest shown uncropped |
| Story / Reel | 1080 × 1920 | 9:16 |
| 1000 × 1500 | 2:3 |
1200×630. Every major platform derives its preview from that, and it is the default here. LinkedIn's own guidance says 1200×627, which is close enough that either works — both are offered.
Because the platforms downscale and re-encode whatever you give them. A 2400×1260 export only makes the file larger for the crawler to fetch; it does not make the preview sharper. Facebook's documented minimum is 600×315 and its recommendation is 1200×630.
No. Both are read and drawn in your browser and nothing is sent to a server.
It shrinks and wraps to fit, down to a floor. Past that the last line is ellipsised rather than allowed to run off the image — which is a hint that the headline is too long for a card read at thumbnail size.
Because a share card appears small, in a feed, next to other people's images, often on a phone in daylight. The check is the same WCAG ratio used for body text: below 4.5:1 the title is hard to read at that size, and the warning is worth more here than it is on a web page.
No — this makes the image. The Open Graph Tag Generator writes the og:image, og:title and og:description tags that point at it.
Not on its own. Facebook and LinkedIn cache the preview, so you need their sharing debugger or post inspector to re-scrape the URL after changing the image.
The card is drawn directly onto a canvas rather than styled as DOM and screenshotted. The screenshot route is one import shorter, but it hands the output's exact pixel dimensions, its font metrics and its gradient rendering to a library reimplementing the browser's own layout engine — and an image whose entire purpose is to be exactly 1200 by 630 is the wrong place for an approximation. Drawing straight to a canvas gives the exact size by construction, and the awkward part is then a pure function rather than a rendering quirk: fitting the text. The title auto-fits by binary search for the largest size at which its wrapped lines fit the available box, rather than stepping down a point at a time — a long title would otherwise re-wrap dozens of times on each keystroke while the preview redraws as you type. Wrapping breaks at spaces, and a single word wider than the whole line — a URL, a compound noun, a run of emoji — is broken mid-word, because text running past the edge of a share card looks broken in a way an awkward break does not. If the text still cannot fit at the minimum size, the last line is ellipsised rather than allowed to overflow. Contrast between the chosen text colour and the background is measured with the WCAG formula the site's contrast checker already uses, and a card below 4.5:1 is flagged: a share card is read at thumbnail size in a crowded feed, so low contrast costs more here than it does on a page. An uploaded logo or background image never leaves the browser. The export is a PNG at one device pixel per image pixel, because every platform re-encodes and downscales the image anyway and a 2× export only makes the upload larger.