Meta tag generator with live previews.

Only a handful of head tags affect how a page appears when it is found or shared. The title element is the strongest on-page signal of what the page is about and usually becomes the blue headline in Google. The meta description is often used as the snippet beneath it. The canonical link tells search engines which URL is the main version when the same content is reachable in several ways.

Open Graph tags (og:title, og:description, og:image, og:url, og:type, og:site_name) control the preview card on Facebook, LinkedIn, Slack, WhatsApp, Discord, iMessage and most other apps that unfurl links. X reads its own twitter:card tag and falls back to the Open Graph values for title, description and image, so twitter:card is usually the only X-specific tag you need.

The meta keywords tag is ignored by Google and has been for many years, so this generator leaves it out. For the bigger picture of what search engines look at, see our explainers on SEO and technical SEO.

Google does not truncate titles at a fixed number of characters. It cuts them when they run out of horizontal space, which on desktop is roughly 600 pixels of 20px Arial. A title full of wide letters such as W and M runs out sooner than one with i, l and t, so a 60-character rule of thumb is only an approximation. This tool measures your title in that font and shows where it would be cut.

Google also rewrites titles it considers unhelpful, for example when they are stuffed with keywords, identical across many pages, or do not match the page. Write a specific, accurate title, put the important words first, and keep the brand at the end so truncation removes the least useful part.

There is no length limit for meta descriptions, but Google shows about two lines on desktop, which is roughly 920 pixels of 14px Arial, and often less on mobile. The meter tracks that width. Treat the first 120 or so characters as the part that must work on its own.

Google often writes its own snippet from the page text when it thinks that matches the search better than your description, especially for long-tail queries. You cannot force it, but a description that accurately summarises the page, written for people rather than as a keyword list, is used far more often. Each page should have its own.

Use an image of 1200×630 pixels, a ratio of 1.91:1. Facebook, LinkedIn, Slack and X's large card all display that shape well. Keep important text and faces away from the edges, since some apps crop to a slightly different ratio, and keep the file well under 5 MB; a compressed JPEG or PNG of a few hundred kilobytes is ideal. The image compressor can get it there.

og:image must be an absolute URL that crawlers can fetch without logging in, preferably over HTTPS. Relative paths, images behind authentication and URLs blocked in robots.txt all produce cards without a picture. Add og:image:alt to describe the image for people using screen readers.

Social networks cache what they scraped the first time someone shared your URL, sometimes for weeks. After changing tags, ask each network to fetch the page again: Facebook's Sharing Debugger and LinkedIn's Post Inspector both have a re-scrape button. For X, preview the link in a new post draft. Slack refreshes unfurls on its own schedule, so test with a URL that has a harmless query string appended.

If a card never appears, check that the tags are in the server-rendered HTML. Most scrapers do not run JavaScript, so tags injected only on the client by a single-page app are invisible to them. Server-side rendering or static generation solves that.

Switch to Analyse HTML, open the page in your browser, view the source, and paste the head or the whole document. The tool reads the title, description, canonical, Open Graph and X tags, loads them into the editor so the previews show the current state, and lists problems: missing tags, duplicate titles or descriptions, relative image URLs, a noindex robots tag and an og:title that differs from the title.

The analysis runs on the text you paste, so it works for staging sites and pages behind a login too. Pair it with the schema markup generator for structured data and the robots.txt generator to confirm crawlers can reach the page at all.

On sites with hundreds of pages, titles and descriptions usually come from templates. That is fine as long as each page's version is unique and specific. A product template such as "{product name}: {key attribute} | {brand}" works; "{brand} | Shop" on every product page produces thousands of duplicate titles that Google will rewrite and that tell searchers nothing.

Put the most distinctive information first, because truncation removes the end. For category and listing pages, include what is listed and where; for articles, the topic and the angle; for local pages, the service and the place. Descriptions should add something the title does not, such as a price range, a delivery promise or what the reader will learn.

Audit periodically. A crawler or Search Console will show duplicate and missing titles, and the Analyse HTML mode here is a quick way to check any single page, including pages behind a login or on staging. Our SEO services team runs the same checks at scale during site migrations, where tags most often go missing.