For the complete documentation index, see llms.txt. Every page on this site is also served as Markdown: append `.md` to any URL, or send `Accept: text/markdown`.
Skip to content
Attensira LogoAttensira
Book a demo

AI engines

See whether Claude names your brand, then change it

Tracked on every plan

Attensira asks Claude your buyers' questions every day through its API and records whether it names you, which pages it cites and who it names instead, each with its sample size. Attensira's default first move there is getting you listed on review sites and listicles, until your own readings point elsewhere. You approve what ships.

Start freeBook a demo

7 days free · card required

How Claude searches and cites, per Anthropic's API docs
When it searches
When the answer depends on current or changing facts, including products and companies that may have changed
When it does not
Stable knowledge is answered from training, with no search and no citation
How many searches
One to three for a simple fact; ten or more to compare several products
What a citation carries
The page's URL and title, and up to 150 characters of its text
Freshness
Each result carries the page's age: when it was last updated
Claude API docs: web search tool (read 5 Oct 2026)Anthropic: Claude can now search the web, 20 Mar 2025, updated 27 May 2025

What does Anthropic document about Claude?

The frame at the top is Anthropic's own account, from the API Attensira asks Claude through. A citation quotes at most 150 characters, so state each fact in one sentence. On our own site, 35% of 104 Anthropic requests in 30 days were Claude-User, fetching a page because someone asked Claude.

Anthropic's three crawlers
Claude-SearchBot
Indexes pages for Claude's search answers
Claude-User
Fetches a page when someone asks Claude a question
ClaudeBot
Collects pages that may be used for training
Anthropic: does Anthropic crawl data from the web (read 4 Oct 2026)

How does Claude decide which brands to name?

Claude searches the web and gives direct citations so a reader can check the source. Anthropic launched web search on 20 March 2025 and made it available on every Claude plan worldwide on 27 May 2025.

Anthropic runs three crawlers. Claude-SearchBot indexes pages for search, Claude-User fetches a page when someone asks Claude a question, and ClaudeBot collects pages that may be used for training. Its help page says blocking Claude-SearchBot or Claude-User may reduce a site's visibility in Claude's search answers.

What it leans on

  • Pages Claude-SearchBot and Claude-User may fetch
  • Pages its web search finds, cited in the answer
  • Pages that state the facts a careful reader can check

How do you get cited by Claude?

What the crew works on first in Claude, in order.

Measured
Before and after each change, on the same prompts
Ships when
You approve it (Draft mode is the default)
Project plan · Claude4 steps
  1. 01

    Allow Claude-SearchBot and Claude-User in robots.txt; decide on ClaudeBot separately.

  2. 02

    Get listed on the review sites and roundups that rank tools in your category.

  3. 03

    State specifics a reader can check: who it is for, what it costs, what it does not do.

  4. 04

    Keep one consistent description of your product across your site and listings.

Chief of staff agentChief of staff · orders the workExample · the order the crew usually works in on Claude

Why does Claude matter for your brand?

  • Claude cites the pages it searched, so you can see what it relied on.
  • It is read through its API, so its answers show what the model says, not a local search page.
  • Attensira's default rule for Claude is getting listed on review sites and listicles, until your own readings show otherwise.

What does one Claude reading record?

Named or not
Whether the answer names your brand, and where in the answer. Across readings that is your mention rate.
Cited pages
Every page the answer cites, yours and everyone else's. Across readings that is your citation rate and the sources that decide the answer.
Who instead
The competitors named in your place, so share of voice is measured on the same answers as yours.
Tone
How the answer talks about you. Sentiment is read from the same stored answer, never a separate guess.
Which engines each plan reads, in the docs

How is Claude read, and what does that change?

There is no consumer reader for Claude, so Attensira asks it through its API. In a second market, the country is stated in the question ("User location: DE") rather than searched from that country, and the reading is labelled that way wherever it appears.

That makes a Claude reading a clean view of what the model says, but not of a local search page. Read it next to ChatGPT and Perplexity, which are read as a person in that country sees them. Our own workspace reads ChatGPT and Perplexity only today, so this page has no Claude numbers of ours.

Reading · Claude4 steps
  1. 01

    Ask

    Your buyer's question, through Anthropic's API: there is no consumer reader for Claude.

  2. 02

    Locate

    In a second market, the country is stated in the question ('User location: DE'), not searched from there.

  3. 03

    Label

    The reading says so wherever it appears: location stated in the question, not a local search.

  4. 04

    Compare

    Read it next to ChatGPT and Perplexity, which are read as a person in that country sees them.

Example · how one Claude reading is taken, per docs tracking/models and tracking/markets

Questions about Claude

Does Attensira track Claude?
Yes. Claude is read every day on every plan, for each prompt you track and in each market you add. Each reading records whether Claude named you, which pages it cited and the tone, and every rate carries its sample size. A change inside the noise is reported as no change.
Does Claude search the web?
Yes. Anthropic launched web search in Claude on 20 March 2025 and made it available on all Claude plans globally on 27 May 2025. Answers that use the web carry direct citations.
Which Anthropic crawler should I allow?
Claude-SearchBot indexes pages for search and Claude-User fetches pages when someone asks Claude a question. Anthropic says blocking either may reduce your visibility in Claude's search answers. ClaudeBot is the training crawler, and blocking it is a separate choice.
Why is Claude read through its API?
There is no consumer reader for Claude. Its answers come from the API, and in a second market the country is added to the question as a line such as 'User location: DE'. That is labelled 'location stated in the question, not a local search' wherever it appears.
What does the crew do to get named in Claude?
It starts from Attensira's default rule for Claude, that review sites and listicles decide its answers, and switches to your own readings once a source has 20 of them. It finds the pages Claude cites for your questions and drafts the listing, pitch or correction for each. Pitches go from your own mailbox, after you approve them.

Whatever it takes to win the answer.

Start freeBook a demo

7 days free · card required