Link Preview Generator: Controlling What Shows Up When You Paste a URL

Most explanations of a link preview generator stop at the image: set og:image, get a nicer card when someone pastes your URL. That's the mechanism, but it misses half the actual problem. The link itself is also something you can control, and once you're generating the preview anyway, pairing it with a branded, trackable link changes what you can learn from — and do with — every share, not just how it looks.

The link and the preview are two separate decisions

When you share a raw URL — yoursite.com/blog/post-slug-here — two things happen independently. The og:image and related tags on that page determine what preview card renders. And the URL itself is whatever the page's actual address happens to be: long, potentially ugly, and giving you zero visibility into who clicked it or where it was clicked from. Most teams solve the first problem and never touch the second, treating the link as just the delivery mechanism for the preview rather than something worth designing on its own.

What a branded sharing link adds

A sharing link is a short, branded URL — on your own domain rather than a generic shortener's — that redirects to the destination while sitting in front of it as its own trackable object. The preview that renders when someone pastes it can still be controlled the same way, through the same og:image and meta tag mechanism, but now the link itself carries information: which link performed better in a specific channel, how many clicks it got, and where those clicks came from, broken down by day, country, device, and referrer.

• A branded domain (go.yoursite.com/x) instead of a generic shortener or the raw long URL
• Click analytics by day, country, device, and referrer on every link
• The same controlled og:image and meta tag preview as a normal URL would have
• One link object you can point at a different destination later without reissuing it

Why this matters more once you're sharing the same content in multiple places

The gap becomes obvious once you're promoting the same piece of content across more than one channel — an email newsletter, a Twitter post, a LinkedIn post, a Slack community — all pointing at the same underlying page. With a raw URL, you can control the preview, but you have no way to tell which channel actually drove the clicks; your analytics only show that someone arrived on the page, not which specific share got them there. With a distinct sharing link per channel, each one gets its own preview and its own click count, so you can see directly that the LinkedIn share outperformed the newsletter link for this particular post, without relying on UTM parameter conventions that someone has to remember to apply consistently every time.

The image still has to be right

None of the tracking value shows up if the preview itself is broken, and a branded sharing link doesn't automatically fix a bad og:image — it just adds a layer on top of whatever preview the destination page already produces. The link needs to redirect cleanly enough that crawlers still see the destination's actual meta tags (or the tags set specifically for that link, if the tool supports per-link overrides), and the image behind it still needs to meet the basics: correct 1200×630px dimensions, an absolute reachable URL, text that survives cropping. A trackable link pointing at a broken preview is still a broken preview — it just also happens to log the click.

Per-link preview overrides

One capability worth specifically looking for is the ability to set a different preview image per sharing link, independent of the destination page's own og:image. This matters when you want to test different creative for the same landing page across different audiences, or when a piece of evergreen content gets reshared for a new campaign and you want the preview to reflect the campaign rather than whatever was live when the page originally published. Without per-link control, you're stuck either editing the page's actual meta tags every time (affecting every share of that page, not just this campaign's) or accepting a mismatch between what the link promises and what the destination page's default preview shows.

What click data actually tells you that a preview alone can't

A well-designed preview image affects whether someone clicks; click analytics tell you whether that effect is actually happening, and for whom. Those are different questions, and conflating them leads teams to iterate on preview design based on gut feeling rather than evidence. A preview that looks great in isolation can still underperform if the headline it uses doesn't match what a specific audience responds to, or if the image reads well on desktop but loses legibility at the smaller size most mobile feeds render it at. Without click data segmented by device, that gap is invisible — the image looks fine every time you personally check it on a laptop.

Referrer and country breakdowns matter for a related reason: they tell you where a piece of content is actually resonating, which often doesn't match where you expected it to. A post you assumed would perform best on LinkedIn might actually get most of its clicks from a Slack community you weren't paying close attention to, and that's only visible if the link driving traffic from each channel is distinct enough to measure separately. A single shared URL posted identically everywhere collapses all of that signal into one undifferentiated number.

A/B testing the preview itself

Once you have both a controlled preview and click data on a link, the natural next step is testing whether a different image or headline actually performs better rather than guessing. Running two image variants behind the same sharing link and comparing click-through with a proper significance read is a meaningfully different exercise than eyeballing which of two designs you personally prefer — it's easy for a team to have strong opinions about which version "looks better" that don't correlate at all with which one a real audience actually clicks more often. This only becomes practical once the link, the preview, and the analytics are already unified in one place; running an A/B test across separate tools for hosting the image, generating the link, and measuring clicks is possible but fragile, since each swap of a variant has to be manually kept in sync across all three.

It's also a lower-stakes way to test than changing a page's actual og:image, since a losing variant only affected the share of traffic it was shown to, not every visitor who's ever going to see that link going forward.

How useopengraph brings these together

useopengraph's Sharing Links are branded, trackable short links with your own domain support and click analytics broken down by day, country, device, and referrer — built specifically so the preview and the tracking aren't two separate tools you're stitching together. Because the same platform generates the OG images, a sharing link can point at a specific rendered image rather than only inheriting whatever the destination page happens to show, and Growth-plan-and-up Split Testing takes this further by running two or more image variants on a single link with a statistical significance read, so you're not just tracking clicks passively but actively finding out which preview gets more of them.