Meta Tag Generator

How to use the generator

1. Write the title and watch the pixels, not the characters The bar under the title fills as you type and the two markers are where the text gets cut — desktop first, mobile second. Ten capital Ws and ten lowercase i's are both ten characters and nowhere near the same width, which is why a character count is the wrong instrument for a limit measured in pixels.
2. Set the canonical properly Absolute, https, trailing slash, and pointing at itself unless you genuinely mean to point elsewhere. og:url is generated from it, so the two cannot drift apart — which they otherwise always do.
3. Leave the robots boxes alone unless you mean it The default output restricts nothing, because that is what most pages want. Tick max-image-preview:large if you want the bigger image in results; tick noindex only when you have decided the page should not exist in search.
4. Write og:title as its own line The checkbox that copies your page title into it is off by default, deliberately. See below for why.
5. Paste an og:image URL and read the measurement The card preview loads the real image and reports its actual dimensions and aspect ratio. This catches the single most common broken-share cause before anyone shares anything.
6. Copy the minimal set Full set is the belt-and-braces version with the Twitter duplicates. Minimal drops the seven tags Open Graph already covers. Both work; minimal is what we ship.

Pixel width is the measurement. Character count is a rule of thumb about it.

"Keep your title under 60 characters" is advice about a pixel limit, worn smooth by repetition until people forgot what it was advice about. Google renders your title in a proportional font, which means an i and a W occupy wildly different amounts of the space available. A title of 65 narrow characters can fit where one of 55 wide characters gets cut, and if you are optimising to a character count you will never know which of the two you wrote.

So this tool measures instead of counting. It draws your text to an off-screen canvas with measureText in Arial at the size each line renders — 20px for titles, 14px for descriptions — and reports the width in pixels. The truncation marker in the preview sits exactly where the string runs out of room.

Two honest caveats, because the numbers deserve them. First, the pixel limits — roughly 580px for a desktop title, roughly 920px across the two lines a mobile title gets, roughly 920px for a desktop description and roughly 680px on mobile — are observed figures, measured by the SEO community against live results and revised whenever Google changes its layout. Google has never published them. Its own documentation on snippets (Google, control your snippets in search results, last updated 20 April 2026) says the opposite of a limit: "There's no limit on how long a meta description can be, but the snippet is truncated in Google Search results as needed, typically to fit the device width." Treat the markers as a good guess about a moving target.

Second, and more important: Google may not use your title at all. Its documentation on title links (Google, influencing your title links in search results, last updated 10 December 2025) lists the <title> element as one source among several — alongside the main visual title, <h1> and other headings, og:title, prominent styled text, anchor text and structured data — and says that "if we've detected an issue on the page, we may try to generate an improved title link from anchors, on-page text, or other sources." Half-empty titles, boilerplate repeated across a site, and titles in a different language from the page content are named as triggers. The preview on this page is the best case: what Google will show if it decides your title is worth showing. Write it well and it usually is.

What Open Graph actually needs, and the og:image mistake everyone makes

Open Graph is the protocol Facebook published in 2010 that the rest of the web quietly adopted — LinkedIn, WhatsApp, Slack, Discord, iMessage and X all read it. One set of tags feeds all of them, which is why this tool's minimal output has one Open Graph block and a single twitter:card line rather than two parallel sets. X reads the og: tags when the twitter: equivalents are absent — that is long-observed behaviour rather than a documented promise, so the full set is there if you would rather be explicit.

The tag that breaks is always og:image. Facebook's own image guidance asks for 1200 × 630 at a 1.91:1 aspect ratio, sets the minimum at 200 × 200, notes that anything under 600 × 315 renders as a small thumbnail rather than a wide card, and caps the file at 8MB. Miss the ratio and the platform crops to fit — usually straight through the middle of your headline text. Miss the tag entirely and your share becomes a bare line of text with no image at all, which is the most common reason a link "looks broken" when someone posts it. There is a second, sneakier version: the first person to share a brand-new URL sometimes sees no image because the platform has not fetched it yet, which is precisely why og:image:width and og:image:height are worth including — they let the platform lay out the card before it has the file.

The tool measures the image you paste and tells you its real dimensions and ratio. It reads those off the loaded image in your browser, so it cannot see the file size — the 8MB cap is the one part you have to check yourself.

og:title and the page title are not the same line

This is the difference the incumbents paper over by filling both fields from one input, and it is the single easiest upgrade on most sites.

A <title> is read by someone scanning nine other blue links. It needs the topic at the front, it usually needs the brand at the end so people know whose result it is, and it is competing on clarity — which is why "Ecommerce SEO Services in Mumbai — Acme Retail Group" is a good title and a dreadful social headline.

An og:title is read by someone scrolling a feed of photographs, holidays and arguments. It has one job: earn a stop. No brand suffix — the platform already shows your domain above the headline. No keyword hedging — nobody is searching, they are scrolling. "Your product pages are indexed. They are just not ranking." is a bad title and a good og:title, and the two coexist happily on one page.

The checkbox that copies one into the other is off by default for exactly this reason. Turn it on for pages where the title genuinely is the best available headline; leave it off for anything you actually want shared.

What this tool cannot do

It writes tags from what you type into it. That is the whole job, and the things it is not doing matter more here than on most tools, because a head block can be perfect in isolation and still be wrong on the page.

• It cannot read your page. No crawl, no fetch, no lookup — it is a browser-side text tool. So it cannot tell you whether the title matches what the page says, whether the canonical points at a URL that returns 200, or whether the description you just wrote is already on eleven other pages of the same site.
• It cannot see the tags already in your head. This is the failure that costs real traffic and no generator can catch it: a theme emits a title, an SEO plugin emits a second one, a page builder adds a third, and Google picks whichever it likes. Paste the block, then view source on the live page and check that each tag appears exactly once. If you use an SEO plugin, put these values into the plugin's fields rather than into the template — two plugins fighting over one head is how you get a canonical pointing at your homepage.
• The pixel limits are observed, not published. Covered above and worth repeating. They move when Google's layout moves, and Google may rewrite the title regardless.
• The og:image check is dimensions only. It reads width and height off the image once your browser has loaded it. It cannot see file size, cannot tell you whether the image is hotlink-protected against a specific crawler, and if the image fails to load here it may still be perfectly visible to a platform — in which case the tool says "could not load" rather than pretending to a verdict it does not have.
• It does not do structured data. Meta tags and schema are different layers solving different problems. For JSON-LD, use the schema markup generator.

One deliberate omission worth naming: there is no field for <meta name="keywords">, and there will not be one. Google's documentation says it has no effect on indexing or ranking at all, and every generator that still asks for it is training people to publish their keyword list in public.

Where to go next

Tags are the last five minutes of on-page work, not the first. Wording the description itself is its own craft, covered in how to write meta descriptions. What you put in the rest comes from on-page SEO — the topic the page is genuinely about, the query it answers, the words your reader uses. What the result looks like once it ranks depends on SERP features, which decide whether your snippet sits above an AI Overview, below a product carousel or inside a sitelink block. And if two URLs are fighting over one piece of content, no title will fix it — that is a job for canonical tags. The complete SEO guide is the map, and the other free tools cover the neighbouring jobs.