Safety Browser agent by Anthropic
Is Claude in Chrome safe?
Short answer Anthropic uses classifiers that block or pause flagged actions, but says the risk is not zero and recommends avoiding banking, health records and other sensitive sites. support.claude.com · checked
2 of 6 safety questions answered on Anthropic's own pages , Asks before sensitive actions: Depends on your settings , Not published: sees your passwords, trains on your data, where it runs (isolation), how long data is kept
What Anthropic states about connecting your accounts. AideDex reports the vendor's published statements, each linked to its source; it does not test or audit them. "Not published" means we found no statement, not that the answer is no. Facts checked up to .
- Status Available
- Where 185 countries (Available in 185 countries, not in Hong Kong)
- Plan $20/mo (Cheapest plan that includes it: Pro)
The six questions
2 of 6 answered-
What data can Claude in Chrome access?
Works inside your own Chrome profile, so it can read and act on the sites you are signed in to. You grant access site by site; Team and Enterprise admins can allowlist or blocklist sites. claude.com · checked -
Does Claude in Chrome ask before sending, buying or deleting?
Depends on your settingssupport.claude.com · checkedThree modes: Manually approve; Automatically approve (default in the Cowork side panel per the permissions guide; can be switched off); or Skip all approvals. Full quote
Three modes: Manually approve; Automatically approve (default in the Cowork side panel per the permissions guide; can be switched off); or Skip all approvals. In every mode it may not make purchases, create accounts, handle card or ID data, download from untrusted sources or permanently delete things. -
Can Claude in Chrome see my passwords?
Not published Anthropic does not state this on the pages we checked. -
Does Claude in Chrome use my data to train AI models?
Not published Anthropic does not state this on the pages we checked. -
Where does Claude in Chrome run, and is it isolated?
Not published Anthropic does not state this on the pages we checked. -
How long does Claude in Chrome keep my data?
Not published Anthropic does not state this on the pages we checked.
Other things Anthropic states
2 statements- Anthropic says safety classifiers block or pause flagged actions but "The risk is not zero": a new prompt-injection attack could lead to data exfiltration. It advises skipping banking, health records and sites with credentials you wouldn't share. support.claude.com · checked
- Model training on Free, Pro and Max is "Opt-out"; Team and Enterprise have no training on your content by default. claude.com · checked
More questions about Claude in Chrome and your data
1 answer- Is Claude in Chrome safe from prompt injection?
- Anthropic uses classifiers that block or pause flagged actions, but says the risk is not zero and recommends avoiding banking, health records and other sensitive sites. support.claude.com · checked
Compare and read on
- Is Perplexity Comet safe? Claude in Chrome vs Perplexity Comet
- Is Opera Neon safe? Claude in Chrome vs Opera Neon
- Is Claude Cowork safe? Claude Cowork vs Claude in Chrome
All alternatives to Claude in Chrome · Is it safe to connect Gmail to an AI agent? What each vendor says it does · How facts are checked