Claude-User
Fetches a page in real time when a Claude user asks Claude to open or summarize a specific URL.
| Operator | Anthropic |
|---|---|
| Powers | Live page fetches when Claude acts for a user |
| Purpose | User-triggered fetch |
| User-agent token | Claude-User |
| Respects robots.txt | Yes |
Claude-User is not a bulk crawler. It fires when a person using Claude triggers a fetch — for example by asking Claude to read or summarize a specific page.
Because the request is initiated by a human in the loop, many sites choose to allow Claude-User even when they block the ClaudeBot training crawler. Anthropic previously used the Claude-Web and anthropic-ai tokens for similar access; Claude-User and Claude-SearchBot are the current tokens to set rules for.
Full user-agent string
Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Claude-User/1.0; [email protected])
Allow Claude-User
Users explicitly asking Claude about your page can actually read it — a direct, intent-driven visit.
User-agent: Claude-User Allow: /
Block Claude-User
You want to prevent all programmatic access from Claude, including user-initiated fetches.
User-agent: Claude-User Disallow: /
Can Claude-User read your page right now?
Test any URL and see exactly what AI crawlers receive.