Our article2 min read
Can an AI agent log in for you without seeing your password? 1Password says yes
1Password has launched a way for AI agents to sign in to your accounts while you approve each login and the agent never sees the password. Here is how it works and what is still unproven.
In this article
Key takeaways
- 1Password announced 1Password for Agents on October 22, 2026. An agent asks to log in, you approve, and 1Password signs in for it.
- 1Password says the agent never gets your password. That is the company’s own claim, and the reports we read had no independent testing.
- Banks have started spotting agent activity in customer accounts, and a Chase spokesperson advised customers not to share usernames and passwords with anyone.
- Pricing and availability for the new launch were not stated in TechCrunch’s report.
The problem it tries to solve
AI agents that book, shop or manage accounts for you often need to sign in. The shortcut many people use is to give the agent a username and password, which 1Password says can leave copies of sensitive data outside your control.
TechCrunch notes the launch landed the same day as an Axios report that US banks have begun detecting agent activity in customer accounts and often cannot tell whether it reflects what the customer asked for.
Why sign-in is the sticking point
Agents are being built to shop, book and manage accounts for you, as Walmart’s plan to let customers connect their own agent shows. Each of those jobs eventually reaches a login box. Whoever controls that moment controls how much risk you take on.
That is why a middle layer matters. Instead of trusting each agent with a password, you trust one tool that holds the credentials and asks you first. It does not remove all risk, because you still have to decide which agents and which accounts deserve access.
How 1Password for Agents works
As TechCrunch describes it, when an agent tries to log in it sends 1Password a unique digital key to confirm its identity. You get an alert about the attempt and must approve it. Then 1Password signs the agent in and fills in the credentials itself.
Named agents include Claude, OpenAI’s Dots and ChatGPT Work, Meta’s Muse, Grok Bot and agents from the startup Town. 1Password says it is working with others. TechCrunch does not say which browsers are supported.
This builds on a July launch
The Next Web reported in July that 1Password had added a similar feature for Claude on Mac. It shows which login is requested and why, injects it only after biometric approval, ends access when the task is done and checks whether secrets leaked onto the page.
That report also says whether the approach resists prompt injection, where hidden instructions on a page trick an agent, remains to be seen. The security design is 1Password’s own claim.
What is not known yet
- Pricing and plans for 1Password for Agents.
- Whether you must approve every login or can pre-approve some.
- How it works with banks and other sites with strict identity checks.
- How it performs in practice, since no independent testing was reported.
What to do now
- Avoid pasting passwords into an agent’s chat or settings.
- If you use an agent for accounts, prefer a method where you approve each login yourself.
- Use a separate, low-risk account while you test any agent, and review activity afterwards.
- Turn on two-factor authentication for important accounts.
Go deeper
Sources
Written by an AI from the sources above and checked by a person before publishing. We have not tested products unless we say so. See our editorial policy or tell us about an error through contact.
