Anthropic-AI is the robots.txt user-agent token used by Anthropic, the AI safety company behind the Claude family of large language models, to identify its web crawlers. Anthropic operates at least two crawlers: ClaudeBot, used for training-data collection, and Claude-SearchBot, used for real-time web retrieval when Claude’s web search feature is enabled. Webmasters can block Anthropic’s training crawler by adding User-agent: anthropic-ai with a Disallow directive to their robots.txt file, which signals that the site’s content should not be used to train Claude models. As with other AI crawler opt-outs, blocking the training crawler does not prevent a site from being retrieved and cited by Claude during live searches if it appears in Claude’s retrieval index. Anthropic has committed to respecting robots.txt directives and publishes documentation on its crawler identifiers, in line with the broader industry move toward greater AI crawler transparency following publisher pressure in 2023.