Show up in Claude. It reads the open web.

It can, when web search is on. Claude's help page on web search says Claude calls a search tool when a question needs current information, and adds: "Every response includes citations".

Anthropic's documentation of the search tool lists what triggers a search. One item is information about specific organizations that might have changed. A question about who is open, what something costs or who serves a town is that kind of question.

Without a search, Claude answers from what it learned in training, which has a cut-off date and may not include a small business at all.

Anthropic's subprocessor list names Brave Search, TurboPuffer and Parallel Web Systems under web search. Anthropic does not say how the work is split between them. TechCrunch reported in March 2025 that Claude's citations matched Brave's results.

Anthropic also runs its own crawlers, described in the next step. We found no documentation of a listings database such as the ones Google and Yelp keep.

Anthropic documents three crawlers and says all three follow robots.txt.

• Claude-SearchBot indexes pages to improve search results.
• Claude-User fetches a page when a person asks Claude about it.
• ClaudeBot collects training data.

Open yoursite.com/robots.txt. If it blocks the first two, or blocks every crawler, Anthropic says the site's visibility in Claude's search answers may be reduced. Blocking ClaudeBot is a separate choice about training. Some hosts turn bots away before they reach the file, so also look for a setting about AI crawlers in your website builder or security service.

Since Brave Search is one of the named providers, check it directly.

1. Go to search.brave.com and search for your business name and city.
2. Look for your own page in the results, not only directory listings.
3. If it is missing, check robots.txt again, then look at who links to you. A link from a local directory, a supplier or a chamber of commerce is a path for a crawler to follow.

That last point is our inference from how crawlers work. Brave does not publish a rule for when a small site is indexed.

Claude quotes pages, so the page has to contain the sentence. Write the services, the city and the areas you serve, the hours, the phone number and the prices you publish as text. The full list is in what a page needs to be quoted.

Show when the page was last updated. The search tool passes each result's page age to the model, described as "when the site was last updated". Nobody documents how Claude weighs it. A visible, current date costs nothing.

Name the place. The same tool can localize results by city, region and country. A page that says "Bend, Oregon" in its first lines can match a search that carries that place.

Ask Claude the question a customer would ask, with your city in it, in a new chat with web search on, at least five times. Then read the citations under each answer. They show which pages Claude used, which is more than most assistants tell you.

If a competitor is named, open the pages cited for them. What do those pages say that yours does not?

An OnMap page covers steps 1 and 3: it is open to Anthropic's crawlers, and the facts are plain text with the date they were last updated. In our own traffic, ClaudeBot made 326 of the 867 requests that Cloudflare counted from AI crawlers on OnMap.me in the week to November 15, 2026. That shows the pages are being read. It does not show that any business is being recommended.