Claude-User
Operated by Anthropic
Assistant fetch
Intermediary access
What it does
Retrieves a page because a user asked Claude about it. Distinct from ClaudeBot, which crawls for training.
How to identify it
- User agent
-
Claude-User - Request signing
- No Web Bot Auth directory published.
- IP ranges
- https://claude.com/crawling/bots.json
- robots.txt
- Respected uniformly, including for user-triggered fetches.
- First seen
- 2025-03
Worth knowing
Anthropic is the outlier here: it applies robots.txt uniformly to all of its bots including the user-triggered one, where Google, OpenAI and Perplexity all carve out user-initiated fetches. Anthropic publishes the bare token only, not a full user agent string, so strings circulating in bot directories are community observations rather than documented values. No Web Bot Auth directory is published at anthropic.com or claude.com.