Claude visibility

Get your brand named and cited by Claude

When Claude searches the web, it relies on a search index plus Anthropic's own page fetches. Anthropic lists Brave Search as its web-search provider, and Brave follows your Googlebot rules. Here is how Claude finds sources, what to allow, and how StarRise tracks it.

Get my free report AI crawler guide

How Claude finds sources

A search index, plus Anthropic's own fetches

Claude answers from what it learned in training, and from the web when a question calls for a search. For web search, Anthropic lists Brave Search as its provider, and Claude's citations closely track Brave's results.

So access is decided in three places. Claude-SearchBot indexes pages for Claude's search. Claude-User fetches a page when a user's question needs it. And Brave's crawler doesn't identify itself separately: it follows your Googlebot rules, so a page closed to Googlebot is closed to Brave, and to Claude's web search too.

What Claude knows without searching comes from training, so it can name a brand it has read about elsewhere. A citation, a link to your own page, needs a page it can reach and read today.

At a glance

Claude and your site

Its index: A web-search index (Anthropic lists Brave Search as its provider), plus Anthropic's own page fetches.
Must be allowed to be cited: Claude-SearchBot and Claude-User. Don't block Googlebot, which Brave's crawler follows.
Training only, your choice: ClaudeBot collects training data. Blocking it doesn't remove you from Claude's search answers.

Access checklist

Which bots Claude needs

Older guides tell you to allow ClaudeBot. That bot only affects training. These are the ones that put you in Claude's answers.

Bot	Run by	What it does	Recommended
Claude-SearchBot	Anthropic	Indexes pages for Claude's web search	Allow
Claude-User	Anthropic	Fetches a page when a user's question needs it	Allow
Googlebot rules	Google (followed by Brave)	Brave's crawler has no separate user agent and obeys your Googlebot rules	Allow
ClaudeBot	Anthropic	Collects text to train future models. No effect on today's search results	Your choice

The rule most allowlists miss. Many robots.txt files name the OpenAI and Anthropic bots and stop there. Brave's crawler doesn't announce itself; it follows whatever you set for Googlebot. If a page is closed to Googlebot by a rule, a CDN setting or a “User-agent: *” disallow, Brave skips it too, and Claude's web search has less of your site to work with. Accurate to November 2026; vendors change this often. The AI crawler guide covers every engine.

What StarRise measures on Claude

Claude, in the same view as every other engine

On plans that include Claude, StarRise asks it your buyers' questions and scores the answers the same way as ChatGPT, Gemini and Perplexity, so you can compare engines directly.

How each AI engine sees the brand, engine by engine. Demo data.

Mention rate

How often Claude names you for your category's questions.

Share of voice

How prominently you appear next to the competitors you track.

Verified citations

Only real links to your domain count. The pages Claude cites instead are listed too.

How Claude describes you

Direct Perception shows the words Claude attaches to your name, strengths and weaknesses included.

Engine by engine

Engines often disagree. Seeing Claude next to the others shows where a fix is engine-specific.

Every raw answer

Prompt Explorer shows exactly what Claude said, question by question and run by run.

Pro can add Claude; White Glove includes it. White Glove also adds execution support from our GEO team. See pricing.

How to improve

Six steps to get cited by Claude

Access first, then pages worth quoting, then the wider web Claude learns from. Mentioned is being known. Cited is being trusted.

Step 1

Allow the bots that cite

Allow Claude-SearchBot and Claude-User in robots.txt, don't block Googlebot (Brave follows its rules), and confirm your CDN or firewall doesn't block them.

Step 2

Check that Brave can find you

Search for your brand and your main service terms on Brave Search. If your key pages don't show up, Claude's web search is unlikely to find them either.

Step 3

Answer the buyer's question on the page

One page per important question, with a direct answer in the first paragraph that Claude can quote as written.

Step 4

Back claims with evidence

Measured results, statistics with their source, and named experts. Engines quote evidence more readily than opinion.

Step 5

Be present on the wider web

What Claude knows without searching comes from training: directories, press, LinkedIn and industry articles. Consistent, accurate profiles there shape how it describes you.

Step 6

Measure, then repeat

Fixes on your own site show up as engines re-read it, typically within a few weeks. Wider citation authority builds over a couple of months.

FAQ

Claude visibility questions

Where does Claude get its web results?

From a web-search index plus Anthropic's own page fetches. Anthropic lists Brave Search as its web-search provider, and Claude's citations closely track Brave's results. Brave's crawler follows your Googlebot rules, so keep Googlebot open alongside Claude-SearchBot and Claude-User.

Does blocking ClaudeBot hide me from Claude?

No. ClaudeBot collects training data only. Blocking it keeps your content out of future training and has no effect on today's search answers. Blocking Claude-SearchBot, Claude-User or Googlebot (which Brave follows) is what stops Claude citing you.

Why track Claude as well as ChatGPT?

Buyers don't stick to one assistant, and the engines often disagree about who to recommend. A one-engine number hides that. Tracking Claude shows whether a gap is everywhere or specific to one engine.

Which plan includes Claude?

Pro can add Claude, and White Glove includes it. Free and Starter don't include Claude. See pricing.

How long until Claude starts citing us?

Fixes on your own site show up as engines re-read your pages, typically within a few weeks. Being cited more widely builds over a couple of months. What Claude knows without searching changes only when the model is updated. We don't promise ranks or dates, and every run re-asks the same questions so you can see what moved.

Other engines

Each engine reads the web its own way

Each engine has its own index and its own bots. Here is how the others pick sources.