Claude SEO means making your site easy for Anthropic’s Claude to find, cite and describe correctly when it searches the web. Claude has searched the web on every plan since May 2025, cites a source in every search answer, and uses three separate crawlers, only one of which is about training.
Claude is a smaller source of traffic than ChatGPT, but its share of AI traffic has been growing. This guide covers how Claude web search works, how it cites, what its crawlers do and what blocking each one costs, and how to become a source Claude cites. Where Anthropic has said something, it is quoted; where something is only reported, it says so.
What Claude SEO means
Sources: Similarweb[10], BrightEdge[11], Anthropic[3][4]. Anthropic does not publish a user count for the Claude apps.
How Claude got web search
How Claude web search works
When a question needs current information, Claude searches the web, reads the results and writes an answer with citations[2]. Claude can also fetch a specific page a user gives it, pulling the whole article into its context[2].
Claude citations: how Claude cites sources
Anthropic’s help page is direct: “Every response includes citations, so you can easily verify sources yourself”, with direct citations, source links and relevant quotes[2]. For developers using the API, citations are always on for web search; each carries the page URL, title and up to 150 characters of the cited text, and Anthropic requires citations to be shown when API output is displayed to end users[4].
That 150-character limit is a useful discipline for writers. A citation can only quote a short passage, so the sentence you want quoted has to stand on its own: a clear claim, with its number or its name, in one sentence.
ClaudeBot, Claude-SearchBot and Claude-User
Anthropic uses three crawlers, and its documentation, updated in April 2026, gives each its own consequence if you block it[3]. Until early 2026 the documentation talked mainly about ClaudeBot[8], which is why many robots.txt files still treat “Claude” as one thing. Use the planner to see what each choice means:
# All three Anthropic crawlers are allowed. # Nothing needs to be added for them.
What is ClaudeBot, and should you block it?
ClaudeBot is the crawler that turns up most in server logs, which is why “ClaudeBot” is searched far more than “Claude SEO”. It collects web content that could contribute to training Anthropic’s models, and it supports Crawl-delay if you want it to slow down[3].
Whether to block it is a content decision, not a visibility one. In August 2025 Cloudflare measured Anthropic’s crawl-to-refer ratio at nearly 50,000 crawls for every visit referred back, against 887 for OpenAI and 118 for Perplexity[9]. Publishers who see their content as the product often block training crawlers for that reason. For most businesses, whose content exists to be found, the cost of blocking ClaudeBot is low and the benefit is small too. What matters for visibility is keeping Claude-SearchBot and Claude-User open.
For the full list of AI crawlers and what each costs to block, see AI traffic analytics. To write the rules, use the robots.txt generator and tester, which has a “Block AI training only” preset.
How to get cited by Claude
Nobody outside Anthropic knows exactly how Claude chooses what to cite. These steps follow from how its search works and what it does with pages. Your ticks are saved in this browser.
Claude vs ChatGPT vs Perplexity for SEO
| Claude | ChatGPT | Perplexity | |
|---|---|---|---|
| Search source | Not named by Anthropic; Brave reported | Search providers, including Microsoft | Its own crawler and index |
| Training crawler | ClaudeBot | GPTBot | None (says PerplexityBot is not for training) |
| Search crawler | Claude-SearchBot | OAI-SearchBot | PerplexityBot |
| User-triggered fetch honours robots.txt | ✓Yes | May not apply | Generally ignores |
| Citations in every search answer | ✓Yes | ✓Yes | ✓Yes |
| Web search on the free plan | ✓Yes | ✓Yes | ✓Yes |
Visibility does not transfer between them automatically. A brand can be cited by Perplexity and absent from Claude, which is why each engine is worth checking on its own. See the guides to ChatGPT SEO and Google AI Mode for the other two big surfaces.
Measuring traffic from Claude
Claude sends real but modest traffic. Across 74,752 sites using Ahrefs’ analytics in March 2026, AI assistants sent about 3.5 million visits, and this is how they split[12]:
| ChatGPT | 2700 |
|---|---|
| Perplexity | 230 |
| Gemini | 230 |
| Claude | 102 |
The trend matters more than the size. Similarweb reported Claude’s share of generative AI web traffic rising from about 2% in June 2025 to close to 9% in May 2026[10], and BrightEdge put Claude at 2.3% of AI referrals to sites in the first quarter of 2026[11].
In Google Analytics 4, visits from claude.ai arrive as referrals. GA4’s AI Assistant channel, added in May 2026, lists assistants by referrer, but its channel definition does not name Claude[13], so check where claude.ai sessions land in your own reports. Our guide to measuring AI traffic in GA4 covers the setup.
Using Claude for SEO work
“Claude for SEO” also means the other direction: using Claude AI as an SEO assistant. With web search on, it can research a topic, compare what the pages ranking for a query cover, draft a brief or a first draft, suggest headings and FAQ questions, and write structured data from a page you paste in. Two cautions apply to every AI tool, Claude included: check any statistic or quote it gives you against the source it cites, and have someone who knows the subject edit anything you publish. Our AI SEO guide covers which SEO tasks AI helps with and where it needs a person.
Questions people ask
What is Claude SEO?
Claude SEO is optimising a website so that Anthropic's Claude can find it, cite it and describe it accurately when it searches the web to answer a question. It combines crawler access (letting Anthropic's search and user-fetch crawlers in), clear, answer-first content, and presence in the sources Claude's search draws on.
Does Claude search the web?
Yes. Anthropic added web search to Claude on 20 March 2025 for paid users in the United States, and made it available globally on all Claude plans, including free, on 27 May 2025. On Team and Enterprise plans an owner must switch it on first. When it searches, every response includes citations.
What search engine does Claude use?
Anthropic's own help pages do not name one. Brave Search was added to Anthropic's subprocessor list the day before web search launched, and testing showed Claude's citations matching Brave's results, so Brave is widely reported as the provider. Anthropic's help page does say image results are powered by Bing.
What is ClaudeBot?
ClaudeBot is Anthropic's web crawler for collecting content that could contribute to training its models. Blocking it in robots.txt signals that your site's future content should be excluded from training. Anthropic runs two other crawlers: Claude-SearchBot for search indexing and Claude-User for fetching pages when a user asks.
Should I block ClaudeBot?
Only if you want to keep your content out of Anthropic's training data. Anthropic does not say blocking ClaudeBot affects search visibility. Blocking Claude-SearchBot or Claude-User is different: Anthropic says that may reduce your visibility in Claude's answers. To opt out of training but stay in search, block ClaudeBot alone.
Does Claude respect robots.txt?
Anthropic says all three of its crawlers honour robots.txt and support the Crawl-delay extension, and it makes no exception for its user-triggered Claude-User, unlike OpenAI and Perplexity for theirs. Rules have to be set on every subdomain you want to opt out, and Anthropic warns that blocking its IP addresses may not reliably work.
How does Claude cite sources?
In the Claude apps, every web search response includes citations, with direct links to sources and relevant quotes where appropriate. Through the API, citations are always on for web search, and each one carries the source URL, title and up to 150 characters of the cited text.
How do I get my website cited by Claude?
Allow Claude-SearchBot and Claude-User in robots.txt, make sure your pages are indexable by mainstream search engines, answer specific questions directly near the top of each section, include evidence Claude can quote, and earn mentions in the third-party sources that rank for your topic. Then check the questions that matter repeatedly to see whether Claude names you.
What is the difference between Claude-User and Claude-SearchBot?
Claude-SearchBot crawls to improve search results for Claude users, building what Claude knows about pages ahead of time. Claude-User visits a page at the moment a user's question needs it. Blocking either may reduce your visibility in Claude, in different ways.
How can I see traffic from Claude?
Visits from claude.ai show up as referrals in your analytics. Google Analytics 4 added an AI Assistant channel in May 2026; its release note lists Claude among the examples, but its channel definition does not name Claude, so check where claude.ai sessions land in your own reports. Visits from mobile apps may arrive without a referrer and appear as direct.
Sources
Anthropic’s blog, help centre and developer documentation are primary sources. Brave’s role is reported by Simon Willison from Anthropic’s subprocessor list. Cloudflare, Similarweb, BrightEdge and Ahrefs are third-party data with their own methods.