Short answer
Claude in Chrome is Anthropic's browser extension that lets Claude read, click and fill in forms on websites inside a person's own Chrome browser. Anthropic's rules say it may not make purchases, create accounts or handle card data. Anthropic does not document a user agent, IP list or signature for it, so it is not the same as ClaudeBot, Claude-User, Claude-SearchBot or the hosted Claude-Agent browser.
Key takeaways
- Claude in Chrome runs inside a real person's Chrome, with their logins.
- Anthropic says it is prohibited from making purchases or creating accounts.
- Anthropic's robots.txt guidance does not mention Claude in Chrome.
What is Claude in Chrome, and what does it do on a website?
Claude in Chrome is a browser extension from Anthropic. In Anthropic's words, it "allows Claude to read, click, and navigate websites alongside you."
A person installs it in Google Chrome, signs in with their Claude account and gives Claude a task. Anthropic says Claude "can view the page you’re on and take actions like reading and typing text, clicking links, navigating between pages, and filling out forms, using your existing logins."
So when Claude in Chrome visits your site, it is not a crawler collecting pages in bulk. It is one real person's browser, with that person's cookies and sessions, being driven by Claude for a task that person asked for. Typical jobs include comparing products across open tabs, pulling details off a page into a note, or walking through a form step by step.
A short timeline, from Anthropic's own pages:
- 25 August 2025: Anthropic announced a pilot with 1,000 Max plan users.
- 24 November 2025: beta for all Max plan subscribers.
- 18 December 2025: available to Pro, Team and Enterprise plans.
- 26 August 2026: Anthropic wrote that "Claude in Chrome is now generally available on every paid Claude plan."
At the time of writing, Anthropic says it does not run on other Chromium-based browsers or on mobile.
Can Claude in Chrome buy things from my store?
Anthropic's permissions guide says no. It lists actions Claude is prohibited from taking "regardless of permissions". They include:
- "Making purchases or financial transactions"
- "Creating accounts"
- "Handling sensitive credit card or ID data"
Anthropic's safety page adds "Bypassing captchas" to the list of things Claude is prohibited from.
That means a shopper may use Claude in Chrome to research your products, read your shipping and returns policies or compare you with a competitor. Under Anthropic's rules, the purchase itself is left to the person.
Anthropic also places responsibility on the user. Its safety page says: "You remain responsible for all browser actions taken by Claude performed on your behalf." The list that follows includes "Purchases or financial transactions" and "Respecting third-party website terms of service, including any restrictions on automated access". So Anthropic's pages are not fully consistent: the permissions guide prohibits purchases, while the safety page still names them among the user's responsibilities. Either way, the person stays responsible for what is bought.
Does Claude in Chrome ask the user before it acts?
It depends on the mode the user picks. Anthropic describes three:
| Mode | What Anthropic says happens |
|---|---|
| Manually approve | Claude pauses and asks before each action. |
| Automatically approve | Claude keeps working, reviews each action for safety and blocks anything it judges unsafe. This is the default in the side panel. |
| Skip all approvals | Claude does not pause, and nothing checks its actions automatically. |
In every mode, Anthropic says Claude needs explicit permission before entering potentially sensitive information, granting authorizations, changing permission settings or making permanent deletions. The prohibited actions above stay blocked whatever the mode.
The automatic mode relies on classifiers. In the general availability announcement, Anthropic says: "A classifier reviews actions Claude is about to take, such as navigating to a new website or entering text into a page, and checks them against what you originally asked for. If the action doesn’t match your request, it’s blocked."
How does Claude in Chrome identify itself to my website?
Anthropic does not document a way. We found no user agent token, IP address list or request signature for Claude in Chrome in Anthropic's announcements or help center.
The extension runs inside the person's own Chrome, using their logins. Anthropic does not describe any signal that sets those requests apart from that person's ordinary browsing. One permission the extension asks for "lets the extension identify itself to Anthropic's servers", but that is about Anthropic's servers, not yours.
Do not confuse Claude in Chrome with Claude-Agent. Anthropic documents Claude-Agent for a different setup: "Some Claude products, including Claude Duet, can use a browser hosted by Anthropic to complete tasks at a user's direction." Those requests carry the browser's normal user agent with a Claude-Agent token added, and come from IP addresses Anthropic publishes at https://claude.com/.well-known/sail.
Anthropic's own warning applies to any token: "Anyone can send the same User-Agent header, so be sure to check that the traffic originated from one of the listed IPs before concluding that it is genuine." For more on why a user agent alone proves little, see Know Your Agent (KYA).
How is Claude in Chrome different from ClaudeBot, Claude-User and Claude-SearchBot?
They do different jobs, and Anthropic documents them separately. Anthropic says agentic browser use "is different from Anthropic's robots (crawlers and fetchers, including ClaudeBot, Claude-User, and Claude-SearchBot)."
| Name | What it is | How Anthropic says to identify it |
|---|---|---|
| ClaudeBot | Collects web content that could contribute to model training. | User agent token, Anthropic's published IPs. |
| Claude-SearchBot | Crawls to improve search results for Claude users. | User agent token, published IPs. |
| Claude-User | Fetches pages when a person asks Claude a question. | User agent token, published IPs. |
| Claude-Agent | A browser hosted by Anthropic doing a task for a user. | Claude-Agent token plus IPs in the SAIL file. |
| Claude in Chrome | An extension driving the user's own Chrome. | Not documented. |
Your robots.txt rules for ClaudeBot, Claude-User and Claude-SearchBot are covered in Anthropic's crawler guidance. That guidance does not mention Claude in Chrome. For a full list of the robots.txt names, see robots.txt for AI agents.
What should a store owner do about Claude in Chrome?
Treat it as a customer using a tool, because that is what Anthropic describes. A few practical steps:
- Keep product pages readable. Clear names, prices, stock status, shipping costs and return terms help a person and the agent reading the page for them.
- Expect the person to finish checkout. Anthropic says Claude may not make purchases, so you do not need to redesign checkout for it.
- Do not hide instructions for agents in your pages. Anthropic describes hidden instructions in web content as prompt injection. It says probes scan web content, and when one detects a likely attack, Claude is warned to treat the content with suspicion.
- Do not rely on detecting it. Without a documented signal, rules that try to single out Claude in Chrome rest on guesswork.
- Write down your terms. If you restrict automated access, say so in your terms of service. Anthropic tells users that respecting those terms is their responsibility.
For a different approach, compare ChatGPT agent, which signs its requests so a site can verify them with Web Bot Auth.
How Sill helps. Sill shows you which AI agents and bots visit your site and the pages they read, and it verifies signed agent requests on Cloudflare and WordPress installs. When an AI agent places an order through Sill, your rules approve it, decline it or hold it for a person to review, and every decision is signed and recorded. You can check how your site looks to agents with the free readiness scan.
Questions people ask
Is Claude in Chrome free?
No. Anthropic says it is available on paid plans: Pro, Max, Ultra, Team and Enterprise. It does not run on other Chromium browsers or on mobile.
Does Claude in Chrome take screenshots of my website?
Yes. Anthropic's safety guide says whatever is visible in the tabs Claude is working in is captured in screenshots and becomes part of the user's conversation.
Can a company stop its own staff using Claude in Chrome on certain sites?
Yes, on Team and Enterprise plans. Admins can set allowlists and blocklists for the sites Claude may open. These are controls for the customer's organisation, not for website owners.
Does blocking ClaudeBot in robots.txt stop Claude in Chrome?
Anthropic does not say so. Its robots.txt guidance covers ClaudeBot, Claude-User and Claude-SearchBot, and it does not mention Claude in Chrome.
Sources
- Claude: Claude in Chrome is generally available · checked 16 Nov 2026
- Anthropic: Piloting Claude in Chrome · checked 16 Nov 2026
- Claude Help Center: Get started with Claude in Chrome · checked 16 Nov 2026
- Claude Help Center: Claude in Chrome permissions guide · checked 16 Nov 2026
- Claude Help Center: Use Claude in Chrome safely · checked 16 Nov 2026
- Claude Help Center: Claude in Chrome admin controls · checked 16 Nov 2026
- Claude Help Center: Does Anthropic crawl data from the web, and how can site owners block the crawler? · checked 16 Nov 2026
- Claude Help Center: How do Anthropic-hosted browsers identify themselves when they browse the web on a user's behalf? · checked 16 Nov 2026
- Anthropic: Claude-Agent IP list (SAIL file) · checked 16 Nov 2026