anthropic.com
Grade C
Audited Nov 10
49 checks
SEO, GEO and AEO audit · home page score ·
SEO audit for anthropic.com
anthropic.com scored 71 out of 100 (grade C) in the ratemysite.ai audit of 49 checks across technical SEO, content, answer engines, AI readiness and speed, with 21 issues to fix.
Email me about this score
Compare with another site
Show your score on your site
Audit complete. This site needs work
Summary
Site health score
Grade C71 out of 100


This site needs work. 21 things to improve.
First audit of this domain
Technical SEO
94
14 checks · 1 warning
Content
72
8 checks · 5 warnings
Answer engines
42
9 checks · 3 failed, 2 warnings
AI readiness
71
12 checks · 1 failed, 5 warnings
Speed
68
6 checks · 1 failed, 3 warnings
Fix first
21 of 48 checksWarning: Direct answer after the H1Answer engines (AEO)
first paragraph after the H1 has 18 words, aim for 25 or more
Medium impactHow to fix: Direct answer after the H1
Warning: Title tag lengthTechnical SEO
title is 16 characters, recommended 30 to 60
Medium impactHow to fix: Title tag length
Core Web Vitals
Phone, field dataLargest Contentful Paint: Passed2345 ms
Cumulative Layout Shift: Passed0.00
Interaction to Next Paint: Warning204 ms
Time to First Byte: Warning897 ms
Site health
The other pages were not scanned in this audit.
AI crawler access
16 of 16 allowedGPTBotAllowedOAI-SearchBotAllowedChatGPT-UserAllowedClaudeBotAllowedClaude-SearchBotAllowedClaude-UserAllowed
Google preview
Desktopanthropic.com
Home \ Anthropic
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
Every check, by category
Technical SEO
Technical SEO is in great shape
Whether search engines can crawl, index and understand the page: title, meta description, headings, canonical, HTTPS, robots.txt and sitemap.
Warning: Title tag lengthMedium impact
What we found
title is 16 characters, recommended 30 to 60
Why it matters
A title tag is the page name shown as the link in search results. This check fails without a title and warns when it is under 30 or over 60 characters.
Passed: Meta description
meta description is 121 characters
Passed: Single H1
exactly one H1
Passed: Heading hierarchy
21 headings in order
Passed: Canonical URL
canonical is self-referencing
Passed: Mobile viewport
viewport: width=device-width, initial-scale=1
Passed: HTML lang attribute
lang="en"
Passed: HTTPS and redirect
https with http redirecting to https
Passed: robots.txt
robots.txt present and allows crawling
Passed: XML sitemap
sitemap.xml reachable
Passed: Indexable (no noindex)
page is indexable
Passed: Image alt text
all 5 images have an alt attribute, 5 decorative (alt="")
Passed: Internal links
29 internal links
Content and E-E-A-T
Content and E-E-A-T needs work
Trust and depth signals: a named author, a visible date, contact and privacy pages, enough readable text and sources linked out.
Warning: Contact or about pageLow impact
What we found
no contact or about page linked
Why it matters
A contact page is a page that says how to reach the people behind a site. This check passes when any link's address or text contains contact or about, and warns otherwise.
Warning: Privacy policyLow impact
What we found
no privacy policy linked
Why it matters
A privacy policy is a public page that explains what personal data a site collects and why. This check passes when any link's address or text contains privacy, and warns otherwise.
Warning: ReadabilityLow impact
What we found
Flesch reading ease 46, aim for 50 or more
Why it matters
Flesch reading ease is a score from 0 to 100 that rates how easy text is to read. This check scores the whole page: 50 points or more passes, and lower gets a warning.
Warning: Visible dateLow impact
What we found
no visible published or updated date
Why it matters
A visible date is a line that shows when a page was updated. The check passes on a time tag, date meta tag, structured data date, or text like Updated 2026. Otherwise it warns.
Passed: Content length
316 words of visible text
Passed: Links to sources
2 external domains linked: claude.ai, platform.claude.com
Passed: Title differs from H1
title "home \ anthropic" differs from H1 "ai research and products that put safety at the frontier. ai will have a vast impact on the world. anthropic is a public benefit corporation dedicated to securing its benefits and mitigating its risks."
Keyword consistency
The most used words on the page and where they appear.
| Keyword | Title | Meta description | Headings | On the page |
|---|---|---|---|---|
| claude | No | No | Yes | 11 |
| anthropic | Yes | Yes | Yes | 5 |
| policy | No | No | Yes | 5 |
| fable | No | No | Yes | 4 |
| opus | No | No | Yes | 4 |
| read | No | No | No | 4 |
| research | No | Yes | Yes | 4 |
| sonnet | No | No | Yes | 4 |
| work | No | No | No | 4 |
| announcement | No | No | No | 3 |
Answer engines (AEO)
Answer engines (AEO) has serious issues
How easily answer engines can lift an answer from the page: a direct answer up top, question headings, definitions and structured data.
Warning: Direct answer after the H1Medium impact
What we found
first paragraph after the H1 has 18 words, aim for 25 or more
Why it matters
A direct answer is a short paragraph right under the H1 that answers the page's question. This check fails under 10 words, warns from 10 to 24, and passes at 25 words or more.
Failed: FAQ schemaMedium impact
What we found
no FAQ section or FAQPage schema
Why it matters
FAQPage is a Schema.org type that marks up questions and answers as JSON-LD. This check passes with FAQPage markup, warns on an FAQ heading without it, and fails with neither.
Failed: Organization and WebSite schemaMedium impact
What we found
no Organization, WebSite or Article structured data
Why it matters
Core schema is the structured data that says who runs a site. This check needs 2 of 7 types, such as Organization and WebSite: 1 gets a warning and 0 fails.
Warning: Question-style headingsLow impact
What we found
no H2 or H3 phrased as a question
Why it matters
A question heading is an H2 or H3 written as the question its section answers. The check passes if one ends in a question mark or starts with a word like how or what.
Passed: Lists and tables
1 lists, 0 tables in the main content
Passed: Data points
7 data points (2026, 2026, 40%)
Passed: Definitions
definition sentence found: "Anthropic is a public benefit corporation dedicated to securing its benefits and mitigating its risks"
Not applicable: HowTo schema
no step list on the page
AI readiness (GEO)
AI readiness (GEO) needs work
Whether AI assistants such as ChatGPT, Claude and Perplexity can reach, read and cite the site: crawler access, llms.txt and content that works without JavaScript.
Failed: llms.txtHigh impact
What we found
no /llms.txt
Why it matters
llms.txt is a plain Markdown file at the root of your site. It tells AI assistants what your site is about and which pages matter most. Writing one takes about 10 minutes.
Warning: Brand name consistencyLow impact
What we found
no og:site_name or WebSite/Organization name; AI engines cannot confirm the brand name
Why it matters
Brand consistency means spelling your site name the same way everywhere. This check compares your title tag, og:site_name and Organization schema. Matching names help search engines and AI assistants link every mention to you.
Warning: ai.txtLow impact
What we found
no /.well-known/ai.txt
Why it matters
ai.txt is a plain text file that states how AI systems may use your content. It lives at /.well-known/ai.txt. It is optional and support is uneven, but some crawlers and agents read it.
Warning: Citable entity linksLow impact
What we found
no sameAs links (Wikipedia, LinkedIn, Crunchbase) in structured data
Why it matters
sameAs links are profile URLs listed in your Organization schema, such as LinkedIn or Wikidata. They let search engines and AI assistants confirm who you are. This check looks for at least 1.
Warning: llms-full.txtLow impact
What we found
no /llms-full.txt (optional, gives AI the full content in one file)
Why it matters
llms-full.txt is an optional companion to llms.txt that holds the full text of your key pages. AI assistants can load it in one request. This check looks for it at your site root.
Warning: Markdown versionLow impact
What we found
no Markdown version via Accept: text/markdown (optional)
Why it matters
A Markdown version is a plain text copy of a page that your server returns when an AI agent asks for text/markdown. This check sends 1 such request and looks for a Markdown reply.
Passed: AI crawler access
all tier 1 and tier 2 AI crawlers allowed
Passed: Content without JavaScript
316 words available without JavaScript
Passed: Cache headers
Cache-Control "s-maxage=150, stale-while-revalidate=31535850"
Passed: No paywall or interstitial
no paywall or interstitial detected
Passed: X-Robots-Tag for AI
no AI restriction in X-Robots-Tag
Passed: HTML size
HTML is 172 KB
Speed
Speed needs work
How fast the page loads and responds on a phone, from Google PageSpeed and real Chrome user data when Google has it.
Failed: Page weightMedium impact
What we found
| Metric | This page | Good | Source |
|---|---|---|---|
| Page weight | 4113 KB | 1465 KB or less | PageSpeed field data |
Why it matters
Page weight is the total size of every file a page downloads, such as images, scripts and fonts. This check passes at 1.5 MB or less, warns up to 3 MB and fails above that.
Warning: Interaction to Next PaintLow impact
What we found
| Metric | This page | Good | Source |
|---|---|---|---|
| Interaction to Next Paint | 204 ms | 200 ms or less | PageSpeed field data |
Why it matters
Interaction to Next Paint (INP) is a metric for how fast a page reacts to taps and clicks. This check passes at 200 ms or less, warns up to 500 ms and fails above that.
Warning: Time to First ByteLow impact
What we found
| Metric | This page | Good | Source |
|---|---|---|---|
| Time to First Byte | 897 ms | 800 ms or less | PageSpeed field data |
Why it matters
Time to First Byte (TTFB) is the wait before the first byte of HTML arrives from your server. This check passes at 800 ms or less, warns up to 1,800 ms and fails above that.
Warning: Request countLow impact
What we found
| Metric | This page | Good | Source |
|---|---|---|---|
| Request count | 88 requests | 50 requests or less | PageSpeed field data |
Why it matters
Request count is the number of files a page fetches, such as images, scripts and fonts. This check passes at 50 requests or fewer, warns up to 100 and fails above 100.
Passed: Largest Contentful Paint
Largest Contentful Paint 2345 ms (field data)
Passed: Cumulative Layout Shift
Cumulative Layout Shift 0.00 (field data)
Site details
Technology
Software and libraries we could identify from the page and its headers.
- Next.js(Framework)
- Cloudflare(CDN)
- HSTS(Security)
Social
Profiles linked from the page, and tracking pixels.
- FacebookNot linked
- InstagramNot linked
- X (Twitter)x.com/AnthropicAI
- LinkedInlinkedin.com/company/anthropicresearch
- YouTubeyoutube.com/@anthropic-ai
- TikTokNot linked
- PinterestNot linked
- Tracking pixelsNone found
Local SEO
Only matters for businesses with a place customers visit.
- Phone on the pageNot found
- Address on the pageNot found
- LocalBusiness schemaNot found
- Google Maps or Business Profile linkNot found
Site insights
What we found on the page beyond the checks. None of it changes the score.
AI crawler access
Which AI crawlers your robots.txt lets in, and the rule that decides.
16 of 16 allowed, 0 blocked, 0 partly blocked
GPTBotOpenAIAllowedCollects content to train OpenAI models
Allow: /(applies to all crawlers)OAI-SearchBotOpenAIAllowedIndexes pages for ChatGPT search answers
Allow: /(applies to all crawlers)ChatGPT-UserOpenAIAllowedFetches a page when a ChatGPT user asks for it
Allow: /(applies to all crawlers)ClaudeBotAnthropicAllowedCollects content to train Claude models
Allow: /(applies to all crawlers)Claude-SearchBotAnthropicAllowedIndexes pages for Claude search answers
Allow: /(applies to all crawlers)Claude-UserAnthropicAllowedFetches a page when a Claude user asks for it
Allow: /(applies to all crawlers)PerplexityBotPerplexityAllowedIndexes pages for Perplexity answers
Allow: /(applies to all crawlers)Perplexity-UserPerplexityAllowedFetches a page when a Perplexity user asks for it
Allow: /(applies to all crawlers)Google-ExtendedGoogleAllowedA control token, not a crawler: tells Google whether Gemini may use your content
Allow: /(applies to all crawlers)Applebot-ExtendedAppleAllowedA control token, not a crawler: tells Apple whether its AI models may use your content
Allow: /(applies to all crawlers)AmazonbotAmazonAllowedIndexes pages for Alexa and Amazon answers
Allow: /(applies to all crawlers)meta-externalagentMetaAllowedCollects content for Meta AI
Allow: /(applies to all crawlers)CCBotCommon CrawlAllowedBuilds the open web dataset many AI models train on
Allow: /(applies to all crawlers)BytespiderByteDanceAllowedCollects content for ByteDance AI models
Allow: /(applies to all crawlers)DuckAssistBotDuckDuckGoAllowedFetches pages for DuckDuckGo AI answers
Allow: /(applies to all crawlers)MistralAI-UserMistralAllowedFetches a page when a Le Chat user asks for it
Allow: /(applies to all crawlers)
ChatGPT-User, Claude-User, Perplexity-User and MistralAI-User fetch pages when a person asks, and may not follow robots.txt rules.
Google search preview
An approximate preview of how this page can look in Google results.
anthropic.com
Home \ Anthropic
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
anthropic.com
Home \ Anthropic
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
Google PageSpeed
Google's lab test of this page on a phone and on a desktop.
| Metric | Phone | Desktop |
|---|---|---|
| Google score | 67 | 90 |
| First Contentful Paint | 1.7 s | 361 ms |
| Speed Index | 6.1 s | 1.6 s |
| Largest Contentful Paint | 9.6 s | 1.8 s |
| Time to Interactive | 17.7 s | 3.2 s |
| Total Blocking Time | 142 ms | 84 ms |
| Cumulative Layout Shift | 0.00 | 0.00 |
Phone opportunities
- Render-blocking requestsEst savings of 1,520 ms
- Reduce unused JavaScriptEst savings of 536 KiB, saves about 750 ms
- Reduce unused CSSEst savings of 91 KiB, saves about 450 ms
- Use efficient cache lifetimesEst savings of 44 KiB
- Improve image deliveryEst savings of 405 KiB
- Legacy JavaScriptEst savings of 14 KiB
- Avoid enormous network payloadsTotal size was 4,113 KiB
- Minify JavaScriptEst savings of 21 KiB
Desktop opportunities
- Render-blocking requestsEst savings of 270 ms
- Reduce unused JavaScriptEst savings of 537 KiB, saves about 200 ms
- Reduce unused CSSEst savings of 89 KiB, saves about 100 ms
- Use efficient cache lifetimesEst savings of 44 KiB
- Improve image deliveryEst savings of 125 KiB
- Legacy JavaScriptEst savings of 14 KiB
- Avoid enormous network payloadsTotal size was 3,801 KiB
- Minify JavaScriptEst savings of 21 KiB
Page weight (phone)
| Type | Requests | Size |
|---|---|---|
| Total | 88 | 4.0 MB |
| Script | 53 | 1.5 MB |
| Media | 2 | 1024 KB |
| Font | 7 | 875 KB |
| Image | 6 | 501 KB |
| Stylesheet | 12 | 159 KB |
| Document | 2 | 65 KB |
| Other | 6 | 4 KB |
JavaScript errors (desktop)
8 errors in the browser console.
- security: Connecting to 'https://analytics-ipv6.tiktokw.us/ipv6/enrich_ipv6' violates the following Content Security Policy directive: "connect-src 'self' https://api.segment.io https://*.segment.io https://*.s
- security: Connecting to 'https://analytics.google.com/g/collect?v=2&tid=G-MT0111843S>m=45je6b61v9195589781za20ozb9195660759zd9195660759xf1&_p=1794279696322&_gaz=1&gcs=G111&gcd=13v3v3v3v5l1&npa=0&dma=0&_eu=AAA
- security: Connecting to 'https://stats.g.doubleclick.net/g/collect?v=2&ngs=1&ibt=1&tid=G-MT0111843S&cid=952763548.1794279697>m=45je6b61v9195589781za20ozb9195660759zd9195660759xf1&rcb=11&aip=1&dma=0&gcs=G111&g
- security: Connecting to 'https://www.google.com/g/collect?v=2&tid=G-MT0111843S>m=45je6b61v9195589781za20ozb9195660759zd9195660759xf1&_p=1794279696322&_gaz=1&gcs=G111&gcd=13v3v3v3v5l1&npa=0&dma=0&_eu=AAAAACAC&
- javascript: Fetch API cannot load https://analytics.google.com/g/collect?v=2&tid=G-MT0111843S>m=45je6b61v9195589781za20ozb9195660759zd9195660759xf1&_p=1794279696322&_gaz=1&gcs=G111&gcd=13v3v3v3v5l1&npa=0&dma=0&
- javascript: Fetch API cannot load https://www.google.com/g/collect?v=2&tid=G-MT0111843S>m=45je6b61v9195589781za20ozb9195660759zd9195660759xf1&_p=1794279696322&_gaz=1&gcs=G111&gcd=13v3v3v3v5l1&npa=0&dma=0&_eu=AA
- security: Loading the script 'https://js.hs-banner.com/v2/23987127/banner.js' violates the following Content Security Policy directive: "script-src 'self' 'unsafe-inline' https://cdn.segment.com https://a-cdn.a
- security: Loading the script 'https://js.hsadspixel.net/pixels.js' violates the following Content Security Policy directive: "script-src 'self' 'unsafe-inline' https://cdn.segment.com https://a-cdn.anthropic.co
Headings
H1 1, H2 2, H3 18
- H1AI research and products that put safety at the frontier. AI will have a vast impact on the world. Anthropic is a public
- H2Claude Fable 5.2 and Mythos 5.2
- H3Introducing Fable 5.2 and Mythos 5.2
- H3Introducing Claude Opus 5.5
- H3Introducing Claude Sonnet 5.5
- H2Learn moreLearn more
- H3The Anthropic Institute
- H3Claude's Constitution
- H3Responsible scaling policy
- H3Core views on AI safety
- H3Research
- H3Policy
- H3Commitments
- H3Company
- H3Products
- H3Models
- H3Solutions
- H3Help and security
- H3Terms and policies
- H3Our mission
- H3Log in
Links on this page
35 links: 29 internal, 6 external, 3 nofollow.
- Read more https://www.anthropic.com/claude-fable-and-mythos-5-2Vague link text
Email security
SPF and DMARC records stop others from sending email that pretends to come from this domain.
- SPFFound: other servers are marked as suspicious (soft fail)
v=spf1 include:_spf.google.com ~all - DMARCFound: fake emails are rejected
v=DMARC1; p=reject; sp=reject; rua=mailto:f0cef134490d4d43b9c134fc3c2a3481@dmarc-reports.cloudflare.net,mailto:dmarc-reports@anthropic.com; ruf=mailto:dmarc-reports@anthropic.com; fo=1at _dmarc.anthropic.com
Score history
The overall score of each audit of this site, oldest first.
First audit of this domain
Questions about this report
What is the SEO score of anthropic.com?
anthropic.com scored 71 out of 100, grade C, with technical SEO 94, content 72, answer engines 42, AI readiness 71 and speed 68.
What should anthropic.com fix first?
Start with llms.txt: no /llms.txt. This report has 21 issues to fix in total.
How is the site health score calculated?
Site health is a score from 0 to 100 that combines five category scores. Technical SEO counts for 25%, content counts for 15%, answer engines counts for 20%, AI readiness counts for 25% and speed counts for 15%.
How often can this report be updated?
A new audit can run 24 hours after the last one. To run it, use the Re-analyze button at the top of this page.
Is this audit free?
Yes. Audits are free and need no account.