Anthropic's training crawler. It collects public pages that may be used to train Claude models. Intent: training. Downloads your content to train AI models.

What is the ClaudeBot user agent?

Requests from ClaudeBot carry "claudebot" in the User-Agent header (matching is case-insensitive). The full string includes a version number that changes over time; Anthropic's documentation lists the current one. In robots.txt, the token is ClaudeBot.

How do I verify ClaudeBot?

Anthropic publishes the IP ranges ClaudeBot uses. A request is real when its IP is inside that list. Paste an IP into HadoSEO's free bot verifier to check it against the published data.

How do I block ClaudeBot in robots.txt?

Add "User-agent: ClaudeBot" followed by "Disallow: /" to your robots.txt. Blocking ClaudeBot does not block Claude-SearchBot or Claude-User: each has its own token and its own rule. robots.txt is a request, not a lock: a scraper using ClaudeBot's name ignores it, which is why verification matters.

Where does Anthropic publish ClaudeBot's IP addresses?

At https://claude.com/crawling/bots.json. The older docs.anthropic.com/claudebot.json address now redirects to an error page, so tools still pointed at it load no ranges and can report the real ClaudeBot as fake. The same file covers ClaudeBot, Claude-SearchBot and Claude-User.

Can I see when ClaudeBot visits my site?

Yes. HadoSEO logs every ClaudeBot request by page, checks it against Anthropic's published data, and flags requests that use the name from IPs Anthropic doesn't own. For a one-off look, the free log file analyzer reads a week of your server logs in the browser.

[ AI training crawler · Anthropic ]

ClaudeBot: Anthropic's AI training crawler.

Anthropic's training crawler. It collects public pages that may be used to train Claude models. Its intent is training: downloads your content to train AI models.

trainingverifiable by published IP ranges

IP ranges last updated Oct 25, 2026

Check a ClaudeBot IP All AI crawlers

[ technical facts ]

operator: Anthropic
intent: training
user agent contains: claudebot
robots.txt token: ClaudeBot
verification: published IP rangesverified
official IP list: claude.com/crawling/bots.json
ranges in registry: none loaded (source unreachable at last refresh)
last updated: Oct 25, 2026
operator docs: support.anthropic.com/en/articles/8896518

[ verify ]

How to verify a real ClaudeBot request.

Anthropic publishes the IP ranges ClaudeBot uses. A request is real when its IP is inside that list.

01Check the User-Agent header contains "claudebot" (case-insensitive). That is only a claim: anyone can send it.
02Download Anthropic's published list (https://claude.com/crawling/bots.json) and check the request IP falls inside one of its ranges.
03Inside the list: the real ClaudeBot. Outside it: a request using ClaudeBot's name from someone else's server.

Skip the lookups: the free bot verifier checks one IP against the same registry, and the log file analyzer checks every ClaudeBot request in a week of access logs. Requests that fail are flagged as spoofed, not hidden.

[ robots.txt ]

Block or allow ClaudeBot in robots.txt.

Add one of these groups to the robots.txt at your domain root. The token matches case-insensitively. Anthropic runs other crawlers with their own tokens, so a rule for ClaudeBot leaves Claude-SearchBot (search index) and Claude-User (answer fetch) untouched.

block ClaudeBot

User-agent: ClaudeBot
Disallow: /

allow ClaudeBot

User-agent: ClaudeBot
Allow: /

robots.txt only binds crawlers that choose to follow it. A scraper sending ClaudeBot's user agent from its own servers won't, which is why HadoSEO checks the IP behind every request.

[ questions ]

ClaudeBot, answered.

[ free scan ]