Claude-Web
AI training crawler operated by Anthropic
Anthropic
Operator
See operator's docs
Verification
Claude-Web
robots.txt token
Yes
Honours robots.txt
What Claude-Web is
Legacy Anthropic crawler user agent.
Its robots.txt token is Claude-Web , and it is listed as honouring robots.txt. Operator documentation: Anthropic .
Verifying it yourself
WebDecoy has no published IP list or reverse DNS hostname on file for Claude-Web. Check Anthropic's documentation for one, such as a list of addresses or signed requests. Until you can check the source address, the User-Agent is only a claim, and copying a User-Agent costs nothing.
Blocking it
If you want the real crawler to stay away, robots.txt is the polite way to ask:
User-agent: Claude-Web
Disallow: /robots.txt only works on a client willing to obey it, which rules out every request that was already lying about who it is. Stopping those takes enforcement at the network or application layer, keyed on something the client cannot simply retype. That is the problem WebDecoy exists to solve.
Frequently Asked Questions
What is Claude-Web? +
Claude-Web is an ai training crawler operated by Anthropic. Legacy Anthropic crawler user agent.
How do I verify Claude-Web? +
WebDecoy has no published IP list or reverse DNS hostname on file for Claude-Web. Check Anthropic's documentation for one; until you can check the source address, the User-Agent is only a claim.
How do I block Claude-Web? +
Add a robots.txt rule for the token Claude-Web. That stops the real crawler, but not clients that only borrow its name, so pair it with verification and enforcement at your edge.
Related crawlers
- Anthropic Anthropic
- Claude-SearchBot Anthropic
- Claude-User Anthropic
- ClaudeBot Anthropic
- AbliterantBot Abliterant
- AI2Bot Allen Institute for AI
- Amazonbot Amazon
- Applebot-Extended Apple
- ByteSpider ByteDance
- CCBot Common Crawl
- ChatGLM-Spider Zhipu AI
- Cohere Cohere
See which Claude-Web requests are real
WebDecoy verifies declared crawlers against what their operators publish and identifies the ones that fail, as persistent actors rather than disposable IPs.