# Claude-Agent

> Claude-Agent is Anthropic's agent that browses and acts on websites on a person's behalf. Here's how to recognize it and check a request really is it.

## Claude-Agent facts

- **Operator:** Anthropic
- **Type:** Declared agent
- **Why it fetches:** Acts in a browser for a person
- **User agent contains:** `Claude-Agent`
- **Verified by:** published IP ranges
- **Operator's documentation:** [support.claude.com/en/articles/18234133](https://support.claude.com/en/articles/18234133)

## Allow or block Claude-Agent

Usually, yes, if you want people's agents to compare, book or buy on your site. Its visits are people's errands: watch what it does before you decide.

To ask it to stay out, add this to robots.txt: `User-agent: Claude-Agent` then `Disallow: /`. Whether it follows robots.txt is Anthropic's call; [its documentation](https://support.claude.com/en/articles/18234133) says.

## How to tell a real Claude-Agent from an impostor

Anyone can put "Claude-Agent" in a user agent. What can't be faked is below.

- **Check the address.** Anthropic publishes the IP ranges Claude-Agent uses: [claude.com/.well-known/sail](https://claude.com/.well-known/sail). A request whose user agent says Claude-Agent from an address outside them is an impostor.

## How Punchline counts Claude-Agent

Punchline labels a request that names Claude-Agent with how sure it is: **verified** or **declared** when the evidence checks out, a lower tier when it can't be checked, and **impostor** when the evidence contradicts it. Each visit is counted under Anthropic, with its purpose, so you can see whether Claude-Agent is bringing people's questions to your pages and which pages it reads. [How it works](https://punchlineanalytics.com/how-it-works) · [Known agents](https://punchlineanalytics.com/docs/agents)

[See which agents read your site](https://app.punchlineanalytics.com/signup)
