Skip to content
AI news · AI tools

1Password now lets AI agents log in for you without seeing your password, like giving the tradie a gate code instead of the house keys

By , Founder of Adaptd ·

1Password for Agents, announced on 22 October, lets AI assistants sign in to your accounts after you approve each request on your own device. The agent never sees or stores the password.

The short version

  • 1Password for Agents lets an AI agent sign in to an account without ever receiving or seeing the password.
  • 1Password announced it on 22 October with partners such as Anthropic and OpenAI, and integrations are rolling out one by one.
  • If you plan to let an AI assistant work inside your accounts, keep the logins in a password manager rather than pasting them into a chat.

AI assistants are getting good at doing jobs inside your accounts, which raises an awkward question: whose password are they using? 1Password's answer is that they shouldn't see one at all.

What is 1Password for Agents?

1Password for Agents is a new 1Password platform that lets AI agents sign in to websites and apps on your behalf without the agent ever seeing or storing the password. When an agent reaches a sign-in page, it identifies itself to 1Password with its own digital key.

You then see which login the agent wants and why, and you approve or refuse it on your own device. 1Password fills in the details, and the password stays out of the AI model's hands. It's the difference between handing someone your keys and buzzing them through the gate.

Which AI assistants will work with it?

The first AI assistants to work with 1Password for Agents come from Meta (Muse), Anthropic (Claude), OpenAI, SpaceXSI (Grok Bot) and Town, according to 1Password's 22 October announcement. Perplexity and Ramp are among nine other partners named.

Timing varies by product. Anthropic says 1Password sign-ins are coming soon to Claude Duet, its proactive assistant app, and 1Password hasn't published a single launch date for every integration.

Should your business let AI agents sign in to its accounts?

Your business should only let AI agents sign in to accounts where you can see and approve each request, and this kind of setup makes that much easier. Pasting a password into a chat box is a habit worth breaking now.

My take: start with low-stakes accounts, like a supplier portal, and keep banking and payroll for humans for a while yet. If your team already uses 1Password, ask your admin to keep an eye on when your preferred assistant switches on.

Trust the agent with the job, not the master key.

Wondering what this means for your business?

That's exactly what the discovery call is for.