Meta tags in 2026: what Google uses and what it ignores

Part of the DMA marketing glossary

Meta tags explained from Google's own docs: the tags Google reads, the ones it ignores, and the 2026 snippet controls for AI Overviews and AI Mode.

JSJun Sing Tan Updated Oct 12, 202619 min readReviewed by DMA editorial team

What you'll learn

  • What are meta tags?
  • Which meta tags does Google support in 2026?
  • Title tags and how Google rewrites title links
  • Meta descriptions and how Google builds snippets
  • Robots meta tag, X-Robots-Tag and data-nosnippet
  • Meta tags and Google's AI Overviews and AI Mode

Meta tags are the short lines of code in a page's head that tell Google what a page is about, how to show it in results and whether to show it at all. Most meta tag guides copy each other, and a few of their rules have been wrong for years. We rebuilt this one on 12 October 2026 from Google's own Search Central documentation, so every claim about what Google does with a tag links to the page where Google says it. You'll get the tags Google supports, the ones it ignores, the new controls for AI Overviews and AI Mode, and code you can paste.

What are meta tags?

Meta tags are HTML elements in the <head> of a page that describe it to search engines and browsers, and to social apps when someone shares the link. People reading the page never see them. Google supports a short list, such as the description and robots tags, and ignores any it doesn't support.

Two head elements that matter just as much aren't technically meta tags: the <title> element and the canonical link. Every guide covers them together, ours included, because they live in the same block of code and describe the page to machines in the same way. Here's a stripped-down head with the basics in place.

HTML head section showing where meta tags are placed in a web page's source code
<!DOCTYPE html>
<html lang="en-SG">
<head>
  <meta charset="UTF-8">
  <meta name="viewport" content="width=device-width, initial-scale=1">
  <title>Aircon servicing in Tampines, same-day slots | ExampleCo</title>
  <meta name="description" content="Same-day aircon servicing in Tampines. Open daily 9am to 9pm.">
  <link rel="canonical" href="https://www.example.com.sg/aircon-servicing-tampines">
</head>
<body>...</body>
</html>

Everything goes between <head> and </head>, which is where Google expects to find it. Google's robots spec adds one quirk: it will still respect a robots meta tag placed in the body, though there is no good reason to put one there (Google robots meta tag spec).

80%+of title links came from the HTML title, Google said in August 2021
1200 x 630pixels, the og:image size Meta recommends
7 May 2026the day Google stopped showing FAQ dropdowns
31 Aug 2026Search Console's AI features switch reached every site

Which meta tags does Google support in 2026?

Google keeps an official list, and it is shorter than most people expect. According to Google's page on the meta tags it supports, these are the ones it reads.

Meta tagWhat Google does with itDo you need it?
descriptionMay use it for the snippet under your resultYes, on every page you want clicks on
robots or googlebotControls indexing and what Google is allowed to showOnly if you want something other than the default
viewportTells Google the page is mobile-friendlyYes, on every template
charset or Content-TypeDeclares the character encoding, and Google advises UTF-8Yes, once per page
google-site-verificationProves to Search Console that you own the siteOn the home page, if you verify that way
notranslateStops Google offering to translate your title link and snippetRarely
nopagereadaloudStops Google's text-to-speech services reading the page aloudRarely
refreshSends visitors to another URL after a delayNo, Google advises a 301 redirect instead
ratingLabels adult content so SafeSearch can filter itOnly when the page has adult content

Two lines on that page deserve a sticky note on your monitor. Google ignores meta tags it doesn't support, so an invented tag costs nothing and does nothing. Google also asks you to avoid injecting or changing meta tags with JavaScript wherever you can, which matters if your site runs on a JavaScript framework.

Need help with your SEO? See our SEO service and prices.

Book a call ›

Title tags and how Google rewrites title links

Google calls the clickable headline of a search result a title link, and it says this is often the main thing people use to decide which result to click. The <title> element is the first source on its list, but Google also reads the main visible heading, <h1> elements and the og:title tag. When those don't settle it, Google can fall back on other large text on the page, the anchor text of links pointing to it and WebSite structured data (Google on title links).

When Google launched its current title system in August 2021, it said the HTML title was still used more than 80% of the time (Google Search Central blog). You'll see other rewrite percentages in 2026 articles. Those come from third-party samples, and we couldn't find a newer figure from Google itself. What Google does publish is the list of problems that make it reach for something other than your title.

Why Google rewrites the titleWhat it looks likeThe fix
Half-empty title<title>| Site Name</title>Write the page's own words before the brand
Obsolete titleThe title says last year while the heading on the page says this yearUpdate the title whenever the page changes
Inaccurate titleThe title promises products the page doesn't showDescribe what is actually on the page
Repeated boilerplateA whole set of pages share one titlePut the detail that differs, like a season or model, in each title
No clear main headingSeveral headings carry the same size and weightMake one heading clearly the most prominent
Wrong language or scriptA Chinese page that kept the English title from the templateWrite the title in the language of the page
Brand named twiceThe site name repeats what Google already showsName the brand once, at the start or the end

The language row is the one Singapore businesses trip on. If your Chinese or Malay page carries the English title from your theme, Google says it may choose a different title link that matches the language of the page.

On length, Google says there's no limit on how long a <title> can be, and it cuts the title link to fit the device. So the 60 character rule you see everywhere is a rule of thumb for dodging that cut; Shopify's help centre uses the same 60 characters for the same reason. Lead with the words a searcher is scanning for. Keep the brand at the end. Google also warns against keyword stuffing, which it says can make a result look spammy.

<title>Aircon servicing in Tampines, same-day slots | ExampleCo</title>

After you change a title, Google has to recrawl the page before the result updates, and it says that can take a few days to a few weeks.

Meta descriptions and how Google builds snippets

The meta description is the summary you write for the snippet under your title. Google says it builds snippets mostly from the page content and uses your description when it describes the page better than the text does (Google on snippets). Snippets also follow the search, so one page can show different text for different queries.

Descriptions have no length limit in Google's docs either, and the snippet is trimmed to fit the screen. Google's advice on writing them is practical:

  • Give each page its own description, and if time is short, start with the home page and your most visited pages.
  • Add facts the snippet would otherwise miss, like a price, your opening hours or an author and date.
  • On a large catalogue site, generate descriptions from your product data; Google encourages it.
  • Leave out strings of keywords, since Google says those are less likely to be shown.

Here's a description for a Singapore service page that gives a searcher a reason to pick it:

<meta name="description" content="Same-day aircon servicing across Singapore. Chemical wash, gas top-up and repairs, open daily 9am to 9pm. Book a slot on WhatsApp.">

Don't expect a description to lift rankings. Google said back in 2009 that while it sometimes uses the description for snippets, it doesn't use it in ranking (Google, September 2009). Its job is to win the click from someone comparing five results on a phone.

Robots meta tag, X-Robots-Tag and data-nosnippet

The robots meta tag tells search engines what they may do with a page. The default is index, follow, and Google's docs say you don't need to write the defaults out, so you only add the tag when you want something different. Use name="robots" for every search engine or name="googlebot" when the rule is for Google alone.

<!-- keep a thank-you page out of search -->
<meta name="robots" content="noindex">

<!-- Google only: show a snippet of up to 160 characters -->
<meta name="googlebot" content="max-snippet:160">

These are the rules that Google lists as valid in its robots meta tag specification, with the cases where a small business would use each.

RuleWhat it does in GoogleWhen you'd use it
noindexKeeps the page out of search resultsThank-you pages, internal search results and staging copies
nofollowGoogle won't follow the links on the pageRarely on a whole page
noneWorks the same as noindex, nofollowPages that should disappear from search entirely
nosnippetNo text snippet or video preview, and no direct input to AI Overviews or AI ModeAlmost never on a page you want customers to find
max-snippet:[number]Caps the snippet length, and also caps how much text AI Overviews and AI Mode can useWhen you want a short preview, while -1 means no limit
max-image-previewSets the largest image preview, from none to standard to largeUse large if you want the bigger thumbnail
notranslateGoogle won't offer a translated title link and snippetPages where a machine translation would mislead
noimageindexImages on the page aren't indexedPhotos you don't want turning up in image search
unavailable_after:[date]Drops the page from results after that dateA promotion or event page that expires, like a Chinese New Year sale
indexifembeddedLets embedded content be indexed despite noindexEmbeds and widgets, and only together with noindex

Three things trip people up. When rules conflict, Google applies the more restrictive one, so max-snippet:50 next to nosnippet means no snippet at all. Google can also only obey a noindex it's allowed to crawl: if you block the URL in robots.txt, the tag is never seen, and Google's canonical guide adds that a blocked URL can still be indexed without its content. The third is noarchive, which still appears in plenty of 2026 guides even though Google says it ignores it now that cached links no longer exist.

X-Robots-Tag for PDFs and images

A PDF has no head to hold a meta tag. For files like that, Google reads the X-Robots-Tag HTTP header, which accepts every rule the meta tag does. If you've got an old price list or brochure PDF that still ranks after the prices changed, a header rule on the server takes it out. On Apache it looks like this, and our technical SEO team can set it up if your hosting panel won't let you.

<Files ~ "\.pdf$">
  Header set X-Robots-Tag "noindex, nofollow"
</Files>

data-nosnippet for one block on a page

Sometimes you want the page in search but don't want one block quoted in the snippet. Wrap that block in data-nosnippet. Google honours it on div, span or section elements, and it asks you not to add or remove the attribute with JavaScript after the page has loaded.

<div data-nosnippet>Promotion ends 31 December. Corporate rates on request.</div>

Same rules on mobile and desktop

Google indexes and ranks the mobile version of your pages, and its mobile-first indexing guide asks for the same robots meta tags on both versions. A noindex left only on a mobile template can stop Google indexing the page. The same guide asks for equivalent titles and descriptions on mobile and desktop.

One related job sits on links rather than pages. Google asks you to mark paid links with rel="sponsored" and links in comments or forum posts with rel="ugc", and to keep rel="nofollow" for any other link you don't want your site associated with (Google on qualifying outbound links).

Meta tags and Google's AI Overviews and AI Mode

The question Singapore owners keep asking us this year is whether AI search needs its own tags. Google has answered it in writing.

There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary.

Google Search Central, AI features and your website

To be shown as a supporting link in either feature, a page has to be indexed and eligible to show a snippet in normal search. Your existing robots tags double as the AI controls. Google's robots spec now says nosnippet stops a page's content being used as a direct input for AI Overviews and AI Mode, while max-snippet limits how much of it can be used. A noindex takes the page out of both, and data-nosnippet hides one block.

The site-wide switch in Search Console

Since 31 August 2026, every Search Console property has a Search generative AI control under Settings (Search Console Help). You can include your site in AI Overviews, AI Mode and the generative AI features in Discover, exclude it, or inherit the setting of a parent property. Include is the default, so you only touch it if you want out. Google says an exclusion generally takes a few days to apply, and that an excluded site gets no traffic or impressions from those features.

The switch doesn't cover AI training. That's handled by Google-Extended, a robots.txt token for Gemini Apps and Vertex AI, and Google says Google-Extended has no effect on inclusion or ranking in Google Search (Google's crawler list).

Our adviceLeave AI features switched on for any page that sells. If one block shouldn't be quoted, such as a promotion with an end date or terms that need context, wrap it in data-nosnippet instead of switching off the page or the whole site. Excluding your site won't stop people getting an answer: Google says content from other sites will still appear, and it may look a lot like yours.

What you can skip

In May 2026 Google published a guide to optimising for generative AI features with a list of things you can ignore. Google Search doesn't use llms.txt files or other AI text files, though keeping one for other services does no harm. Structured data isn't required for AI features, and there's no special schema for them. Google even says optimising for AI search, from where it sits, "is optimizing for the search experience, and thus still SEO."

So if an agency quotes you a separate AI retainer to add meta tags, ask which tag it plans to add, because Google says there isn't one. That's why our AI SEO work sits inside the SEO plan instead of a second retainer.

To see the effect, open the Generative AI performance report in Search Console, which Google rolled out to every site on 31 August 2026. It counts how often links to your pages appeared in AI Overviews and AI Mode, and you can group it by page or by country, Singapore included (Search Console Help).

Canonical and viewport tags, plus hreflang for multilingual sites

Canonical tag

A canonical link tells Google which URL you prefer when the same content lives at more than one address, as happens with tracking parameters or with http and https versions. Google treats it as a strong signal rather than an order. In Google's canonical guide, a redirect is the strongest signal of all and a sitemap listing is a weak one. Google recommends a self-referencing canonical on the preferred page, written as an absolute URL.

<link rel="canonical" href="https://www.example.com.sg/aircon-servicing-tampines">

Two mistakes come up again and again. People use noindex to pick a canonical inside their own site, which Google advises against because it removes the page from search altogether. Others send mixed signals, like one URL in the sitemap and another in the canonical tag. Our guide to canonical URLs walks through the edge cases.

Viewport tag

Google says the viewport tag tells the browser how to render a page on a mobile device, and that its presence indicates the page is mobile-friendly. Because Google indexes the mobile version first, check that every template carries the tag, and that goes for landing pages built outside your main theme too.

<meta name="viewport" content="width=device-width, initial-scale=1">

hreflang for English and Chinese versions

If you run the same page in English and Chinese, or add Malay or Tamil versions, hreflang tells Google they're one page in different languages (Google on localised versions). Each version must list itself and every other version, with full URLs that include https. Google doesn't use hreflang or the lang attribute to work out a page's language; it detects that from the text.

<link rel="alternate" hreflang="en-SG" href="https://www.example.com.sg/en/">
<link rel="alternate" hreflang="zh-Hans-SG" href="https://www.example.com.sg/zh/">
<link rel="alternate" hreflang="ms-SG" href="https://www.example.com.sg/ms/">
<link rel="alternate" hreflang="x-default" href="https://www.example.com.sg/">

Here zh-Hans-SG means Simplified Chinese for users in Singapore. Google won't accept a country code on its own, because the first code always stands for a language.

Charset

<meta charset="UTF-8"> declares how the page's characters are encoded. Google recommends UTF-8 where possible, which is what you want on any site that mixes English with Chinese characters or Tamil script.

Open Graph tags for social previews

Open Graph tags set the title and image people see when your link is shared on Facebook and on other platforms that read the protocol. The spec at ogp.me requires four properties on every page: og:title, og:type, og:image and og:url. A description is optional, and the spec suggests one or two sentences. Since Google lists og:title among its title link sources, a careless OG title can also end up in search.

For the image, Meta's sharing guide asks for at least 1200 x 630 pixels, with 600 x 315 as the floor for the large preview, a ratio close to 1.91:1 and a file under 8 MB. Add og:image:width and og:image:height, because without them the first person to share your link may see no image while Facebook's crawler fetches it. Facebook caches images by URL, so when you change the picture, give the file a new name and run the link through the Sharing Debugger.

<meta property="og:title" content="Aircon servicing in Tampines, same-day slots">
<meta property="og:type" content="website">
<meta property="og:url" content="https://www.example.com.sg/aircon-servicing-tampines">
<meta property="og:image" content="https://www.example.com.sg/og/aircon.jpg">
<meta property="og:image:width" content="1200">
<meta property="og:image:height" content="630">
<meta property="og:description" content="Same-day slots, open daily 9am to 9pm.">
<meta name="twitter:card" content="summary_large_image">

X has its own twitter:card tags, and we've kept one in the sample. We've left out claims about how X reads them, because X's card documentation wasn't on its developer site when we checked on 12 October 2026.

Tags Google ignores, and a feature it dropped

Google's docs keep a list of tags it no longer uses, because people still ask about them. Here's what's on it, plus tags Google never supported.

  • Meta keywords. Google said in 2009 that it doesn't use the keywords tag in ranking, and its docs now say the tag has no effect on indexing or ranking at all. Our meta keywords guide has the history.
  • The lang attribute, since Google works out a page's language from the text itself.
  • rel="next" and rel="prev", which Google says no longer have any effect on indexing.
  • nositelinkssearchbox. The sitelinks search box it controlled is gone.
  • noarchive and nocache, which Google ignores.
  • Author, revisit-after and similar leftovers, because Google ignores any meta tag that isn't on its supported list.

One correction to plenty of older advice, the earlier version of this page included: the rating meta tag still works. Google supports it to label adult content for SafeSearch.

Structured data isn't a meta tag, but it sits in the head as JSON-LD and gets lumped in with them. Google says it can make pages eligible for what it calls "rich results", that it isn't needed for AI features, and that it should match the visible text. If you added FAQ markup for the dropdowns under your result, note that Google stopped showing the FAQ result feature on 7 May 2026 and removed the documentation in June (Search Central updates log). Our schema markup guide covers the types worth adding.

How to add meta tags on WordPress, Shopify, Wix or custom code

Google's own advice is that on a CMS such as WordPress or Wix you may not be able to edit the head directly, so you look for the CMS's search engine settings instead. Here's where those settings live on the common platforms.

  1. WordPress: install an SEO plugin such as Yoast SEO, which adds SEO title and meta description fields to every post and page.
  2. Shopify: in the admin, go to Online Store, then Pages, open the page and click Edit website SEO in the search engine listing section. The home page title and description sit under Online Store, then Preferences. Shopify's title field takes up to 70 characters (Shopify Help).
  3. Wix: in the editor, open Pages and Menu, click More Actions next to a page and choose SEO basics to set the title tag and meta description. Wix creates a default canonical for each page and reuses your SEO title and description for social shares unless you change them (Wix Help).
  4. Custom code or a JavaScript framework: put the tags in the HTML your server sends. Google recommends avoiding JavaScript to inject or change meta tags where you can, and it asks that JavaScript never changes a canonical that's already in the source.

Whichever route you take, test the result with the URL Inspection tool in Search Console, since it shows the HTML Googlebot actually received rather than what your editor shows you.

How to audit your meta tags in an afternoon

You don't need a paid tool to find the expensive mistakes. Work through these in order.

  1. Run your ten most visited pages through URL Inspection and read the robots tags in the HTML Google received.
  2. Crawl the whole site with a tool like Screaming Frog or Sitebulb, then sort for titles and descriptions that are missing or duplicated. Look hardest for a stray noindex left over from a staging site.
  3. Open the Performance report, switch to Pages and sort by impressions. A page that shows up often but gets a low click-through rate is the first place to rewrite a title or description.
  4. Check the Generative AI performance report, and confirm under Settings that nobody has excluded the site from AI features by accident.
  5. Paste your main URLs into Facebook's Sharing Debugger to see the preview and refresh the cache.
  6. Ask Google to recrawl the pages you changed, then give it a few weeks before you judge the result.

Repeat the audit after a redesign or a plugin swap, since that's when head code most often gets overwritten without anyone noticing.

Frequently asked questions

What are meta tags in SEO?

Meta tags are HTML elements in a page's head that describe the page to search engines and browsers. Google supports a short list, including the description and robots tags, and ignores tags it doesn't support. The title element and canonical link aren't meta tags, but they sit beside them and matter as much.

Which meta tags matter most for SEO?

Start with the title element and meta description, because they shape the result people click. Then check the robots meta tag, since a stray noindex removes a page from Google, and the canonical link, which tells Google which URL you prefer. The viewport tag matters because Google indexes the mobile version of your pages.

Does Google still use the meta keywords tag?

No. Google said in 2009 that it doesn't use the keywords meta tag in ranking, and its current documentation says the tag has no effect on indexing or ranking at all. You can delete it from your templates without any loss in Google.

How long should a title tag and meta description be?

Google sets no length limit for either. It trims both to fit the device, so about 60 characters for a title is a sensible rule of thumb to avoid the cut. For descriptions, put the most useful words first and write each page its own summary.

Do I need special meta tags for AI Overviews or AI Mode?

No. Google says there are no extra requirements or special optimisations for AI Overviews or AI Mode. A page needs to be indexed and eligible for a snippet. Google Search also doesn't use llms.txt files, and there's no special schema for AI features.

Will nosnippet keep my content out of AI Overviews?

Yes. Google's robots spec says nosnippet stops a page's content being used as a direct input for AI Overviews and AI Mode, and it also removes the text snippet in normal results. To hide one block only, use data-nosnippet. To opt the whole site out, use the Search generative AI control in Search Console.

Why does Google show a different title from the one I wrote?

Google builds title links from a list of sources and swaps in other text when your title is half-empty, out of date, inaccurate, repeated across pages or in a different language from the page. Fix the cause and request a recrawl. Google says updates can take a few days to a few weeks.

Get a free meta tag check from DMA

Meta tags take minutes to fix and are easy to break across a whole site in one theme update. They won't put a page on page one by themselves, but one stray noindex can take it off Google entirely. If you'd like a second pair of eyes, ask for a free site audit and we'll check your titles, descriptions, robots rules and canonicals against what Google's documentation says. If AI Overviews are already answering your customers' questions above your listing, we'll show you which of your pages they cite and which they skip, and that work sits inside the same SEO plan.

JS

Jun Sing Tan

Jun Sing Tan is part of the content team at D’Marketing Agency, a Singapore digital marketing agency specialising in SEO, SEM, social media & lead generation. About DMA ›

Get SEO and marketing tips

Share f X in
Book your discovery call

Important Notice: Protect Yourself from Scammers

We have been informed that scammers are impersonating D’Marketing Agency Pte. Ltd. Please note that our official website is / , and phone number is +65 8971 9405

Always verify contacts before engaging and do not share personal information with unverified sources. If you have been scammed or approached by scammers impersonating us, please report the number to the platform you are using. We are not responsible for these fraudulent activities.

For any concerns or to confirm the legitimacy of any communication, contact us directly through our official phone number or website. Your safety and security are our top priorities.

Thank you and stay safe!