Skip to content
Brief Security & Trust ·

1Password adds sign-in for AI agents that never shows them passwords

The new 1Password for Agents fills in logins after the user approves, so the agent never sees the password.

In brief
  • 1Password for Agents lets an AI agent log in only after the user approves on their own device.
  • The report does not say what one approval covers or how long it lasts.

The Verge reported on 22 October that 1Password has launched 1Password for Agents. It lets an AI agent get past a login form without ever handling the password. The agent first proves who it is with a unique digital key. 1Password then alerts the user, naming the account and the reason. If the user agrees, the app enters the login details itself. The user approves on their own device, much like a two-factor check.

The Verge says the feature is designed for popular agents such as Meta Muse, Claude Duet and Grok Bot. 1Password says Browserbase, Perplexity and Ramp are among firms working on support. The report gives no price, no release date and no plan details. It also does not say what one approval covers or how long it lasts.

Hiding the password from the agent is only half of the access question. Aashish Tripathi, 1Password's chief product officer, said an agent should not get permanent access to everything behind an account, so the scope and length of each approval is the real control. Buyers could ask: What does one approval cover? How long does it last? Can it be withdrawn? Are agent sign-ins logged?

A WebPulse Brief: a short report of an important event, written by the WebPulse Newsroom with AI assistance and checked against the reporting below. When there is more to explain, we follow up with a full story. How we use AI.

Reporting: The Verge.