How to Get Your Brand Recommended by Claude

Last updated: October 2026

To get your brand recommended by Claude, make sure Claude can find you when it searches the web and can trust what it finds. Allow Anthropic's Claude-SearchBot and Claude-User crawlers, check that you show up in Brave Search, state plainly what you sell and who it is for, and earn mentions on the third-party sites Claude cites.

Key takeaways🔗

• Claude answers from what it learned in training, and it can also search the web. Anthropic's docs say Claude searches when a question depends on current information, including products that "might have changed."

• Anthropic runs three crawlers: ClaudeBot (training), Claude-User (fetches pages when a user asks) and Claude-SearchBot (search quality). Blocking the last two can reduce your visibility in Claude's answers.

• Anthropic's docs do not name a search provider. In March 2025, TechCrunch reported that Anthropic added Brave Search to its subprocessor list, and tests showed matching results.

• Claude cites its web sources, so the pages it finds about you (yours and other people's) shape what it says.

• TrendlyAI tracks Claude on every plan, so you can see whether Claude names you or a competitor week to week.

How does Claude decide which brands to recommend?🔗

Claude decides which brands to recommend from two inputs: what it learned during training and, when it runs a web search, the pages it finds and cites. Anthropic does not publish a ranking formula for brand recommendations, so the practical goal is to be well represented in both.

Training data gives Claude its general knowledge of your category. If your brand was well covered on the public web before a model's training cutoff, Claude may know you without searching. Anthropic's ClaudeBot crawler collects web content "that could potentially contribute" to training (Anthropic).

Web search fills the gaps. Claude got web search in March 2025, first for paid US users, and Anthropic later said it was "available globally on all Claude plans" (Claude). Anthropic's developer documentation explains when Claude searches. It lists "information about specific organizations, people, or products that might have changed" and "current prices, rates, scores, or statistics" as reasons to search, while stable facts and creative tasks are usually answered directly (Anthropic docs).

That documentation is written for developers using the Claude API, so treat it as a guide to how Claude behaves, not a promise about every answer in the Claude app. Still, it points at what matters for brands: buying questions ("best CRM for a 10-person agency," "is X worth the price") involve products and prices that change, which is exactly when Claude is likely to search.

How does Claude's web search work?🔗

Claude's web search works in a loop: Claude decides a search is needed, writes a query, reads the results and then writes an answer that cites its sources. According to Anthropic's docs, this can repeat several times in one answer, and "citations are always enabled for web search" (Anthropic docs).

Two details from the docs matter for marketers:

• Claude writes its own queries. The example in Anthropic's docs shows Claude turning a question into a short query like "claude shannon birth date." Your pages need to rank for the short, plain queries a buyer's question breaks down into, not for the long question itself.

• Search results include a page age. Each result includes a page_age field, described as "when the site was last updated." Claude can see how fresh a page is, so outdated pricing or old "2023" pages put you at a disadvantage on questions about current options.

Simple factual questions typically use one to three searches, while comparisons "can use 10 or more," per the same docs. A "which tool should I pick" question often triggers several searches, which gives more third-party pages a chance to shape the answer.

Which search engine does Claude use?🔗

Anthropic has not named a search provider in its web search documentation, but credible reporting links Claude's web search to Brave Search. On March 21, 2025, TechCrunch reported that Anthropic had added "Brave Search" to its subprocessor list, the list of vendors that process Claude data (TechCrunch).

Developer Simon Willison found further evidence: a search in Claude returned ten citations that were "an exact match for that search on Brave," and Claude's web search function included a parameter named "BraveSearchParams" (Simon Willison). TechCrunch noted that neither Anthropic nor Brave confirmed the arrangement in its article.

So the fair summary: Brave Search is the best-supported answer, based on Anthropic's own subprocessor disclosure as reported in 2025 and independent tests, but Anthropic has not published details on how it is used today, and that could change.

Why it matters: Brave runs its own index, separate from Google and Bing. A brand that ranks well in Google may still be thin in Brave. That is different from ChatGPT, where Bing has the strongest public evidence (see does ChatGPT use Bing?).

Which Anthropic crawlers should you allow?🔗

You should allow Claude-SearchBot and Claude-User if you want to appear in Claude's answers. Blocking ClaudeBot is a separate decision about training. Here is what Anthropic says about each (Anthropic):

A robots.txt that allows the two answer-related crawlers looks like this:

User-agent: Claude-SearchBot
Allow: /

User-agent: Claude-User
Allow: /

Anthropic also notes that robots.txt rules apply per subdomain, and that blocking by IP address "may not work reliably" and stops its crawlers from reading your robots.txt at all. If you use a CDN or firewall with bot protection, check that it is not blocking these user agents by default. That is a common silent failure.

How do you get recommended by Claude? Seven steps🔗

Getting recommended by Claude comes down to being reachable, being findable in the search index it uses, and being described clearly and often by sources it trusts. Work through these in order.

1. Open the door to Anthropic's crawlers🔗

Check your robots.txt, CDN and firewall settings for Claude-SearchBot and Claude-User, as shown above. This is the only step that can take you from invisible to eligible overnight.

2. Check whether you show up in Brave Search🔗

Search search.brave.com for the short queries your buyers use: your category, "best [category] for [audience]," and "[your brand] vs [competitor]." If you do not appear, work on that before anything else.

Brave's help pages give site owners a few useful facts (Brave Search):

• Brave's crawler does not use its own user agent, and "if a domain or page is not crawlable by Googlebot, then Brave Search's bot will not crawl it either." Do not block Googlebot on pages you want in Brave.

• You can submit a page for re-fetching at search.brave.com/submit-url, which helps after you update pricing or product pages.

• robots.txt does not keep a page out of Brave's index. Use a noindex directive for that, and make sure important pages do not carry one.

3. Say clearly what you sell, who it is for and what it costs🔗

Claude can only recommend you for a use case if some page says you serve it. Put a plain description near the top of your home page and key product pages: the category, the type of customer, the main features, the price or price range, and where you operate. "Payroll software for restaurants with 5 to 50 employees, from $40/month, US only" is far more useful to an AI than "the future of work."

Structure helps. Use question-style headings, short direct answers and comparison tables. Our guide to schema markup for AI citations covers the markup side.

4. Publish honest comparison and alternatives pages🔗

When a buyer asks Claude to compare options, it searches for comparisons. If the only "you vs competitor" page is on a competitor's site, that is the page Claude reads. Write your own fair comparison pages, include where competitors are a better fit, and keep prices dated ("as of October 2026"). A fair page is also more useful to a buyer, and to an AI summarizing options for one.

5. Earn mentions on third-party sites🔗

Claude cites what it finds, and for "best" questions that is often review sites, industry publications, "best of" lists and community threads, not brand sites. Get listed on the review platforms in your category, pitch trade publications, and answer real questions in communities where your buyers ask them. Our content strategy for AI citations guide covers how to prioritize this.

6. Keep key pages fresh🔗

Since Claude's search results include when a page was last updated, stale pages are a weak spot. Review your pricing, product and comparison pages every quarter. Change the content when facts change, and update the visible date when you do. Then ask Brave to re-fetch the page.

7. Track what Claude actually says🔗

Ask Claude the questions your buyers ask, on a fixed schedule, and record which brands it names and which sources it cites. One check is not enough. Answers vary from run to run and change as the web changes.

How do you track whether Claude recommends your brand?🔗

You track Claude recommendations by running the same set of buyer prompts in Claude regularly and recording which brands appear and which pages get cited. You can do it by hand in a spreadsheet, or use a tool.

TrendlyAI tracks Claude on every plan, alongside ChatGPT, Perplexity and Google AI Overviews. It runs weekly scans with one sample per prompt per engine, shows when a competitor is named and you are not, and drafts fixes such as comparison pages, FAQ pages with schema and Reddit answers. Plans start at $19/month with a 7-day free trial, no credit card. It does not run daily checks, and one weekly sample means you should read trends over several weeks, not single results. I track TrendlyAI's own domain this way and publish the score, which is still low, so I am not going to pretend any of this is instant.

The free TrendlyAI check is a quick starting point. It runs 5 prompts on ChatGPT and Perplexity in about 30 seconds with no signup. It does not include Claude, so use it as a baseline for the other two engines and test Claude by hand or in a trial. For more on setting up prompt tracking, see how to track brand mentions across AI engines, and for the full picture across engines, visit our GEO hub.

Frequently asked questions🔗

Does Claude use Google or Bing for web search?🔗

Anthropic's web search docs do not name a provider. In March 2025, TechCrunch reported that Anthropic added Brave Search to its subprocessor list, and independent tests found Claude's citations matched Brave results. We found no public statement from Anthropic that Claude's web search runs on Google or Bing.

Will blocking ClaudeBot stop Claude from recommending me?🔗

Not directly, according to Anthropic. ClaudeBot relates to training. Claude-User and Claude-SearchBot are the crawlers tied to answering user questions and search quality. Blocking those two is what Anthropic says may reduce your visibility in Claude's search results.

How long does it take to start appearing in Claude?🔗

There is no published timeline. If you were blocked, unblocking the crawlers can help quickly once your pages are crawled again. Building third-party mentions and ranking in Brave takes weeks to months. Track weekly and look for trends.

Does Claude always search the web before recommending a product?🔗

No. Anthropic's docs say Claude searches when a question depends on current or changing information, such as product details or prices, and answers directly from training for stable knowledge. Many buying questions do trigger a search, but not every one.

Does TrendlyAI track Claude?🔗

Yes. Claude is one of the four engines TrendlyAI tracks on every plan, with ChatGPT, Perplexity and Google AI Overviews. The free check covers ChatGPT and Perplexity only.

Related reading🔗

• Does ChatGPT use Bing? How ChatGPT search finds sources

• How to get cited by Perplexity

• What is generative engine optimization?