1Password launched 1Password for Agents on 22 October. It targets the moment a personal AI agent hits a login page and stops. Today you either take over and type the password yourself, or paste it into the prompt and give the agent more than the task needs.
How it works
- The agent asks for a specific login from your vault.
- You review the request and approve it in 1Password.
- 1Password fills the username, password and one-time code into the site itself, so, in 1Password’s words, the agent does this “without exposing your password or one-time code.”
You can see what is shared and disconnect an integration at any time. When a password changes, connected agents pick up the new one. 1Password adds that “the safeguards and controls vary by integration.”
Who it works with
Not much is live yet. 1Password says integrations for Muse, from Meta, and Grok Bot are “rolling out over the coming days across all platforms.” Anthropic’s Claude Duet, OpenAI’s ChatGPT Work and dots, and Town’s Townie are “coming soon.” Browserbase, Browser Use, KERNEL, Perplexity, Ramp and several smaller agents are listed as coming later. There is a waitlist for agents not on the list.
Price and limits
It costs nothing extra on 1Password Individual and Family accounts on 1Password.com. Accounts on 1Password.eu and 1Password.ca are not supported yet, and availability, platforms and features differ by partner. The blog does not say how other plans are covered.
Why it matters
This week’s Railway breach showed how fast a stored secret spreads once it leaks. An approval step per login is a sensible default. Until your agent’s integration arrives, the safe habit stays the same: do not paste passwords into an agent’s chat.
The Campfire
No commentsNobody has pulled up a log by this one yet. Be the first to say what you make of it.
Held for the desk. It appears after a look.