1Password has launched a tool that lets AI agents sign in to your personal accounts without ever seeing your password. According to TechCrunch AI, the password manager released 1Password for Agents on Thursday. It arrives as more people hand their logins to AI assistants, and as banks start to worry about it.
The idea is simple. Your agent gets access. Your credentials stay locked away.
How It Works
The tool sets up a three-way exchange between the agent, the user and 1Password. Here’s the sequence, as TechCrunch AI describes it:
- The agent identifies itself. When an agent tries to log in for you, it sends a unique digital key to 1Password to confirm who it is. This means the agent has its own verifiable identity instead of borrowing yours.
- You get an alert. 1Password notifies you about the login attempt, and you can approve the request yourself. You stay in charge of each login.
- 1Password fills in the login. Once you approve, 1Password enters the credentials for the agent. The agent gets into the account and finishes its task, but it never holds the password.
That third step matters most. The agent gets a working session, not a copy of your password that it could store and reuse later.
Which Agents It Supports
Users can pick their preferred assistant. The launch list includes:
- Claude Duet from Anthropic
- OpenAI’s Dots and ChatGPT Work
- Grok Bot
- Meta’s Muse
- Agents from Town, an AI assistant startup
1Password says it’s working with other companies to add more. TechCrunch AI’s report doesn’t cover pricing or whether the feature needs a particular 1Password plan, so check the company’s own materials for those details.
The Problem It’s Solving
Right now, most people give an agent account access the lazy way. They paste in their username and password. 1Password says this creates a real risk.
“Handing over a password, secret, or credential is a common shortcut to an agent, but can leave people with sensitive information in places they can no longer control as a copy gets saved for the next time,” the company said in a blog post.
Its pitch: “1Password for Agents gives people a trusted way to delegate tasks to personal AI agents while keeping control of the passwords, secrets, and credentials those tasks require.”
That’s the core issue. Once a password lands in an agent’s memory, a chat log or a third-party server, you no longer control where it goes.
Banks Are Already Uneasy
The timing is pointed. 1Password launched on the same day Axios reported that U.S. banks have begun spotting agent activity inside customer accounts.
That doesn’t mean rogue bots are draining accounts. It mostly reflects consumers asking automated services to handle tasks for them, including financial ones. The trouble is that most banks can’t tell whether a given agent action came from a real user request.
Bankers who spoke with Axios called the situation “horrifying.” A Chase spokesperson told the outlet that customers “should not give their usernames and passwords to anyone.” So banks are telling customers to keep their credentials private, while consumer AI products keep asking for them. That’s the gap 1Password is going after.
Why This Matters
What stands out here is the move toward agent identity. For years, security was built around one question: is this the right human? Agents complicate that. A login might come from you, from your assistant acting on your instructions, or from something pretending to be your assistant.
Giving each agent its own verifiable key, and putting human approval in front of every login, gives services a way to tell those cases apart. It also leaves an audit trail that a shared password never could.
There are caveats, though:
- Approval fatigue is real. If every login needs a tap, some users will start approving without thinking, which weakens the protection.
- The other side still has to adapt. 1Password keeps your password safe, but banks and other services still need their own ways to recognize and trust agent activity.
- Support is limited for now. Coverage depends on which agents 1Password has partnered with, though the company says more are coming.
What Comes Next
Expect rivals to follow quickly. Password managers, identity providers and the AI labs all have reasons to own the layer where agents prove who they are. If agents are going to book travel, pay bills and manage subscriptions for us, “just give it your password” won’t survive long, and the companies that make delegation safe will have a lot of leverage.
You can find more details in TechCrunch AI’s original report.