1Password launches a feature letting AI agents sign in without seeing passwords
Users keep control of approval, while banks report struggling to tell whether agent activity in accounts reflects what customers actually asked for.
Password manager 1Password has released 1Password for Agents, a feature that lets AI agents sign into a user's accounts without ever receiving the password. When an agent reaches a login form, it presents a unique digital key to 1Password, which then notifies the user about which credentials are requested and why. The user approves on their own device, after which 1Password fills in the login details and the agent continues its task. The system is built to work with agents including Meta Muse, ChatGPT Work, Dots, Claude Duet, Grok Bot and Town, with more partners such as Browserbase, Perplexity and Ramp said to be joining. The launch comes as banks report detecting agentic activity in customer accounts and struggle to tell whether it reflects user requests, a concern that Chase raised by telling customers not to share their usernames and passwords with anyone.
Why it matters
The change puts a gatekeeper between AI agents and login credentials. Users can see which account an agent wants to access and why, and must approve each sign-in on their own device, so the agent never holds the password itself. The timing matters because U.S. banks say they are seeing agent activity in customer accounts and often cannot tell whether a user requested it. Chase's warning that customers should not share usernames and passwords with anyone shows how seriously institutions are treating the question of who is acting inside an account.
Matters or hype?
Does this story matter, or is it hype? Vote before you see what everyone else said.