Meta Tags Extractor
Paste any URL to see every meta tag: title, description, canonical, Open Graph, Twitter Card, hreflang, JSON-LD. With 2026 AI-search health-check warnings.
What this tool does (60 seconds)
Paste any URL. We fetch the page server-side and extract every meta tag in the <head>: title, description, canonical, robots, Open Graph (LinkedIn/Facebook/Slack previews), Twitter/X Card, hreflang alternates, JSON-LD structured data, favicons, and viewport. Tags are auto-categorized by purpose. Health-check warnings flag missing or misconfigured tags against 2026 best practices — including AI-search-specific signals like schema content-match and canonical consistency. No login. No JS rendering (we show what crawlers see).
What is a meta tag?
A meta tag is an HTML element inside the <head> of a webpage that provides metadata — information about the page that's meant for browsers, crawlers, and other software, not for human display. Meta tags don't appear visually. They communicate what the page is, who wrote it, how to render it in social previews, and which crawlers may access it.
The category includes several distinct tag families:
- Standard HTML meta tags:
<title>,<meta name="description">,<meta name="robots">,<meta name="viewport"> - Link relation tags:
<link rel="canonical">,<link rel="alternate" hreflang>, favicon links - Open Graph Protocol (ogp.me, originated by Facebook 2010):
og:title,og:image,og:url,og:description - Twitter / X Card tags:
twitter:card,twitter:image - JSON-LD structured data:
<script type="application/ld+json">blocks containing Schema.org markup
The 8 meta tags that matter in 2026
Ranked by combined impact on Google SERP ranking, AI engine citation, and social preview quality. The top 4 are critical for any 2026 SEO/GEO program.
<title>CriticalPurpose: Browser tab title, Google SERP snippet, default OG fallback. The single most-weighted meta tag for both Google and AI engines.
Best practice: 50-60 characters. Include primary keyword near the front. Match (or extend) your H1 — content-mismatch hurts AI citation.
AI signal weight: High — used by ChatGPT, Perplexity, Claude, Gemini to understand page topic during retrieval.
<meta name="description">CriticalPurpose: Google SERP snippet (Google rewrites ~70% of the time per Ahrefs 2023 study), social preview fallback, AI summary context.
Best practice: 140-160 characters. Active verbs, concrete numbers, value proposition. Match the page's actual topic.
AI signal weight: High — AI engines use this as a content summary signal during ranking.
<link rel="canonical">CriticalPurpose: Tells Google (and AI crawlers) which URL is the authoritative version when duplicates exist (UTM-tagged URLs, pagination, parameter variants).
Best practice: Absolute URL pointing to the canonical version. Self-referencing canonical on every page is the safest default.
AI signal weight: Critical — AI engines collapse duplicates via canonical. Wrong canonical = your authority gets routed to someone else's URL.
JSON-LD structured dataCriticalPurpose: Schema.org markup in <script type="application/ld+json">. Article, FAQPage, Organization, BreadcrumbList, Product, HowTo.
Best practice: Schema content-match is what matters — schema fields must mirror rendered page text. Per Ahrefs May 2026 DiD study, schema presence alone is NOT statistically significant for AI citation.
AI signal weight: Critical when content-matched — entity-graph signal for AI engines.
<meta name="robots">HighPurpose: Per-page crawler directives (noindex, nofollow, noimageindex). Different from site-level robots.txt — more granular control.
Best practice: Leave off unless you specifically need to noindex. Default = index, follow. Don't accidentally noindex production pages.
AI signal weight: Medium — AI crawlers respect noindex similarly to Googlebot.
Open Graph tags (og:title, og:image, og:url, og:description)HighPurpose: Social preview cards (LinkedIn, Facebook, Slack, Discord, iMessage). Defined by Open Graph Protocol (ogp.me, 2010).
Best practice: og:image at 1200×630px, under 1 MB, with descriptive filename. og:title can differ from page title for social-optimized framing.
AI signal weight: Low directly — but social shares drive citation indirectly via increased URL exposure.
Twitter / X Card tags (twitter:card, twitter:image, twitter:title)MediumPurpose: X (Twitter) preview cards. Falls back to Open Graph if Twitter-specific tags are missing.
Best practice: twitter:card = summary_large_image for content; summary for product pages. twitter:image min 300×157, max 4096×4096, under 5 MB.
AI signal weight: Low — X preview only.
<link rel="alternate" hreflang="">Medium (Critical for international)Purpose: Tells Google + AI engines which language/region variant to serve based on user locale.
Best practice: Reciprocal hreflang return-tags on every variant. x-default for unmatched locales. Validates via the Google Search Central rules.
AI signal weight: Medium — affects which version AI engines cite for international users.
Are meta keywords still used? (Short answer: no)
Google dropped support for the meta keywords tag in 2009. Matt Cutts announced it on the official Google Search Central blog in September 2009. Bing dropped it years earlier. ChatGPT, Perplexity, Claude, and Gemini don't use it for ranking either.
The tag is harmless but useless. If you see SEO guides recommending meta keywords, that advice is over 15 years out of date. Remove the tag on your next CMS update to reduce unnecessary HTML weight.
Writing good meta tags (2026 patterns)
Title tag patterns
Pattern: {Primary Keyword} — {Modifier} | {Brand} (50-60 chars total). Lead with the keyword users search. Modifier adds context (year, format, audience). Brand suffix on the right gets truncated first in narrow SERPs. Match (or extend) the H1 to maintain content-match for AI citation. Avoid clickbait — AI engines penalize title/content mismatch more aggressively than Google.
Meta description patterns
140-160 characters. Lead with the answer, not the question. Include the primary keyword once. Active verbs (“learn”, “compare”, “build”). One concrete number or specific outcome. End with implicit call-to-action. Google rewrites descriptions ~70% of the time when missing or off-topic (Ahrefs 2023 study) — writing a strong one increases your odds of keeping the version you want.
Open Graph image best practices
1200×630px, under 1 MB. Descriptive filename (“product-name-feature.png”, not “image-1.png”). Text overlay should be readable at thumbnail size. Avoid putting critical info in the bottom-right (often cropped by mobile clients). Test with our /tools/twitter-card-validator for visual preview.
8 common meta tag mistakes (and what fixes them)
Missing or empty <title> tag▾
Every page needs a unique, descriptive <title>. Empty titles produce empty Google SERP snippets and tell AI engines nothing about the page. Most CMSes auto-generate titles from the H1 — verify yours does.
Missing meta description (or duplicate descriptions across pages)▾
Google rewrites descriptions ~70% of the time when missing or off-topic (Ahrefs 2023 study). Write one per page. Duplicate descriptions across multiple pages signal to Google that the pages are templated rather than uniquely useful.
Wrong canonical URL (pointing to a different page or to /)▾
Canonical should be the absolute URL of the current page (or its authoritative version if duplicates exist). A common bug: every page on a CMS has canonical pointing to /. This tells Google your homepage is the canonical version of every URL — catastrophic for indexing.
Open Graph image too small, too large, or missing▾
1200×630px sweet spot. Under 1 MB. Missing og:image = LinkedIn/Slack/Discord show a generic blank card. Reduces click-through dramatically. Use our /tools/twitter-card-validator to preview.
Schema markup that doesn't match page content▾
Per Ahrefs May 2026 DiD study (1,885 pages), schema markup presence alone is NOT statistically significant for AI citation. Content-match is. Empty FAQPage entries (Q with no real answer) hurt more than they help. Either populate schema with real content or omit it.
Using <meta name="keywords"> expecting it to help SEO▾
Google dropped support for the keywords meta tag in 2009 (Matt Cutts blog, September 2009). Bing dropped it earlier. AI engines don't use it. The tag is harmless but useless — remove on next CMS update to reduce HTML weight.
Conflicting robots directives (meta vs robots.txt vs HTTP header)▾
If robots.txt disallows a page, Google can't crawl it to see the meta noindex tag — so the page may still appear in SERPs (without snippet). Use one signal source: either robots.txt OR meta robots, not both for the same URL.
Title and description that don't match rendered H1 / first paragraph▾
Content-match matters for AI citation. If your <title> says "X Pricing" but your H1 says "Y Plans" and your first paragraph talks about Z, AI engines can't determine the topic confidently. Align title, H1, and intro paragraph.
Frequently asked questions
What is a meta tag?▾
A meta tag is an HTML element placed inside the <head> of a webpage that provides metadata about the page — information meant for browsers, search engines, social platforms, and (in 2026) AI engines. Meta tags don't appear visually on the page; they're machine-readable hints about what the page is, who wrote it, how to display it in previews, and which crawlers may access it. Common examples: <title>, <meta name="description">, <link rel="canonical">, Open Graph tags (og:title, og:image), Twitter Card tags, JSON-LD structured data, and hreflang alternates.
Which meta tags actually matter in 2026?▾
Eight tags do the heavy lifting: (1) title, (2) meta description, (3) canonical, (4) JSON-LD structured data, (5) meta robots (only if you need non-default crawler behavior), (6) Open Graph (og:title, og:image, og:url, og:description), (7) Twitter Card tags, (8) hreflang for international sites. Title + description + canonical + JSON-LD are the four critical-priority tags for both Google ranking and AI citation. See the ranked breakdown above for purpose and best practices on each.
Are meta keywords still used for SEO?▾
No. Google dropped support for the meta keywords tag in 2009 (announced on the official Google Search Central blog by Matt Cutts in September 2009). Bing dropped it years earlier. AI engines (ChatGPT, Perplexity, Claude, Gemini) don't use it for ranking either. The tag is harmless but completely useless. If you see SEO guides recommending meta keywords, that advice is 15+ years out of date. Remove the tag on your next CMS update to reduce unnecessary HTML.
What's the difference between meta description and Open Graph description?▾
<meta name="description"> is for Google's SERP snippet (and AI engines as a content summary). og:description is for social previews on LinkedIn, Facebook, Slack, Discord, iMessage. They can be the same text but don't have to be — many sites use a more keyword-focused meta description for SEO and a more emotional/click-driving og:description for social. If og:description is missing, most platforms fall back to <meta name="description">.
Does Google use Open Graph tags for ranking?▾
No. Open Graph tags don't affect Google's ranking algorithm. They affect social media previews (LinkedIn, Facebook, Slack). However, social previews drive click-through on shared URLs — and high click-through-rate URLs tend to earn more backlinks, which indirectly affects ranking. The direct effect of OG tags on ranking is zero; the indirect effect via social sharing is real but hard to measure.
What size should the Open Graph image be?▾
1200×630 pixels is the cross-platform sweet spot. Facebook minimum is 200×200; recommended is 1200×630 (1.91:1 ratio). LinkedIn handles both 1200×627 and 1200×627 well. X (Twitter) summary_large_image: 1200×675 (16:9). Slack and Discord render whatever Facebook accepts. Keep file size under 1 MB (under 5 MB hard cap for X). Use descriptive filenames (e.g., "product-name-feature.png" not "image-1.png").
How does AI search use meta tags?▾
AI engines parse title, description, canonical, and JSON-LD as primary content signals during retrieval and ranking. <title> is used to disambiguate the page topic. Meta description is used as a content summary. Canonical is used to collapse duplicate URLs to a single authoritative version. JSON-LD provides the entity graph (Organization, Person, Article schemas) that AI engines use for source attribution. Per Ahrefs' May 2026 DiD study, schema content-match (schema fields mirroring rendered text) is statistically significant for AI citation, while schema presence alone is not.
Should the title match the H1?▾
Yes — or be a close variant. <title> is the metadata version (often with a brand suffix: "Page Topic — Brand Name"); H1 is the on-page version (just "Page Topic"). Major mismatch (title says one thing, H1 says another) confuses AI engines about the page topic and can hurt both Google ranking and AI citation. The pattern that works: H1 = the on-page headline, title = H1 + brand suffix, meta description = expansion of the H1 with value prop.
Why doesn't the tool render JavaScript?▾
The tool fetches raw HTML server-side. Meta tags injected by client-side JavaScript (after page load) won't appear in our extraction. This is intentional — most search engines and AI crawlers see the raw HTML response, not the post-JS DOM. If your meta tags are JS-injected, fix that with SSR (Next.js, Nuxt, Remix), static generation, or pre-rendering. JS-injected meta tags are invisible to most crawlers and a leading cause of missed indexing.
How is this different from browser DevTools?▾
DevTools shows the post-JavaScript rendered DOM. This tool shows the raw HTML response that crawlers see. For SEO and AI citation purposes, the raw HTML view is what matters. If you have client-side meta tag injection, DevTools will show the tags but Google/AI crawlers may not see them. Use both: DevTools to verify final rendered state; this tool to verify what crawlers receive.
Sources
- Google Search Central blog (Matt Cutts, September 2009) — “Google does not use the keywords meta tag in web ranking.” developers.google.com
- Open Graph Protocol specification. ogp.me
- X / Twitter Cards documentation. developer.x.com
- Google Search Central — title link best practices. developers.google.com
- Google Search Central — meta description best practices. developers.google.com
- Ahrefs 2023 study — Google rewrites meta descriptions ~70% of the time. ahrefs.com/blog
- Ahrefs May 2026 DiD study (1,885 pages) — schema markup presence not statistically significant for AI citation; content-match is. ahrefs.com/blog
- Google Search Central — hreflang documentation. developers.google.com
- Schema.org — structured data vocabulary. schema.org
Related free tools
Want a full AI SEO audit?
This tool checks one thing. TurboAudit runs 250+ checks across 7 branches — schema, E-E-A-T, AI citeability, and more. Start with 5 free audits, no credit card.