Model trainingYour call — not scored
ClaudeBot
Collects content to train Anthropic models. Independent from Claude-SearchBot.
- robots.txt token
ClaudeBotanthropic-aiClaude-Web- Operator
- Anthropic ↗
- Powers
- Anthropic model training
- Honours robots.txt
- Yes, documentedDocumentation ↗
Allow it
User-agent: ClaudeBot
Allow: /Block it
User-agent: ClaudeBot
User-agent: anthropic-ai
User-agent: Claude-Web
Disallow: /Need the whole file, with every crawler and your private paths — and the llms.txt to go with it? The generator writes them. AI config generator →
User-Agent
Mozilla/5.0 (compatible; ClaudeBot/1.0; +claudebot@anthropic.com) Crawlable/1.0; +https://crawlable.fr/probeThis is the string we send when probing. The Crawlable suffix identifies us in your logs.
Anthropic
Check your own site
robots.txt is only half the answer: your CDN can refuse ClaudeBot before it ever reads the file. A scan checks both.
Free, no account, about 5 seconds.