Skip to content
Skimwire
Live wire
AI 2 min skim

1Password now lets AI agents sign in for you without seeing your passwords

1Password for Agents fills in logins for AI agents like Muse and Grok Bot after you approve each request, so the agent never sees the password.

1Password for Agents banner: the 1Password logo beside a grid of AI partner logos on a blue background
Image: 1Password

The skim

  • 1Password launched 1Password for Agents, which lets AI agents sign in to your accounts without ever seeing your password, username or one-time code.
  • When an agent hits a login page, you get a request on your device, approve it, and 1Password fills the login in directly. The agent only learns that access was granted.
  • It's free for 1Password Individual and Family plans on 1Password.com. Meta's Muse and Grok Bot integrations roll out over the coming days. Claude Duet, ChatGPT Work and dots, and Town are "coming soon."

What 1Password announced

AI agents are now booking trips, ordering groceries and filing expense reports, and they keep running into the same wall: a sign-in page. Until now you had two bad options. Take over and type the password yourself, or paste it into the chat and hope it doesn't end up somewhere it shouldn't. "Neither of those are acceptable," 1Password says in its announcement.

Its fix works a lot like two-factor approval. When a connected agent reaches a login, it uses a unique digital key to prove its identity to 1Password. You see which login it wants and why, and you approve or deny it. If you approve, 1Password fills the password or code straight into the sign-in form, and once the site accepts the login, it removes the credential so the agent can't read or reuse it, according to 1Password's FAQ. Connecting an agent doesn't open your whole vault, and access is granted per session for the specific items you approve, Engadget reports.

"Agents are starting to do real work on our behalf. The old model where you hand over a credential and hope for the best doesn't work," 1Password chief product officer Aashish Tripathi said in a statement quoted by The Verge.

Who it works with

The launch partners are Meta's Muse and Grok Bot, with integrations rolling out "over the coming days across all platforms," plus Anthropic's Claude Duet, OpenAI's ChatGPT Work and dots, and Town's Townie, all marked "coming soon." Instinct, Browserbase, Browser Use, Dana, Kernel, Perplexity, Ramp, Tomo and Wajo are next. It costs nothing extra for Individual and Family accounts on 1Password.com. Accounts on 1Password.eu and 1Password.ca aren't supported yet, and features vary by agent.

Why it matters

This week brought a wave of agents that act on your behalf, from Anthropic's Claude Duet to Walmart's agent shopping profiles. Logins are where those agents get stuck, and where things can go badly wrong if a password leaks into a chat log or a model's context. 1Password wants to be the neutral layer every agent uses to sign in, the same role it plays for people today.

It doesn't make agents safe on its own. Once an agent is signed in, it can see and do whatever that account allows, and 1Password itself tells users to review the task and the account before approving. An agent can still order the wrong thing or book the wrong flight. What changes is that your passwords stay out of the agent's hands, and every sign-in needs your OK.

Disclosure: Skimwire is produced with the help of AI tools, including Grok Bot, one of the agents named in this announcement.

Sources: 1Password: Introducing 1Password for Agents, 1Password for Agents FAQ, The Verge, Engadget.

Newsletter

Get the day's tech, skimmed.

Two emails a day, morning and afternoon, with each edition's verified tech stories. Free, unsubscribe anytime.

Check your inbox to confirm.

Something went wrong. Try again.