1Password launched 1Password for Agents on Thursday, a platform that lets personal AI agents use the passwords, secrets and credentials they need for an approved task without ever seeing or storing them. The Toronto-based password manager is launching with a roster of agent partners that includes Meta's Muse, Anthropic's Claude Duet, OpenAI's ChatGPT Work and dots, SpaceXSI's Grok Bot and Town's Townie.
The system is built around a handoff that keeps the secret away from the model. When an agent reaches a sign-in form, it verifies its identity to 1Password with a unique digital key. The user then sees which credential is being requested and why, and approves or declines the request on their own device.
If the user approves, 1Password fills the credential directly into the sign-in form. The agent never receives the password and carries on with its task, while the credential stays in the user's 1Password vault.
Key facts
- 011Password launched 1Password for Agents on Oct. 22, 2026, letting AI agents use credentials without seeing or storing them.
- 02Launch partners include Meta's Muse, Anthropic's Claude Duet, OpenAI's ChatGPT Work and dots, SpaceXSI's Grok Bot and Town's Townie.
- 03Users approve or decline each credential request on their own device before 1Password fills it into the sign-in form.
- 04Browserbase, Browser Use, Perplexity, Ramp and five other companies are also listed as integration partners.
- 051Password says its enterprise vault protects more than 1.6 billion credentials and secrets for over 200,000 businesses.
1Password is aiming the product at the work personal agents have started to take on: planning vacations, managing recurring bills and preparing expense reports. The company said in its announcement that handing an agent a password is a common shortcut, but one that can leave sensitive information copied into places people no longer control.
“Agents are starting to do real work on our behalf.”— Aashish Tripathi, Chief Product Officer, 1Password
Aashish Tripathi, 1Password's chief product officer, said the old approach of handing over a credential and hoping for the best no longer works. An agent, in his framing, should be able to finish the job it was assigned without holding permanent access to everything in the account, and 1Password wants to be the trust layer that makes that possible.
Each launch partner described its own integration. Ami Vora, Anthropic's chief product officer, said Claude Duet asks for permission before it signs in anywhere, and that 1Password will soon complete sign-ins using passwords people have already saved. Glen Coates, head of app platform at OpenAI, said the integration gives agents a way to use accounts users already keep in 1Password without exposing sensitive credentials to the model.
“Most of the work people want help with lives inside their accounts, and we know an agent has to earn its place there”— Ami Vora, Chief Product Officer, Anthropic
Nat Friedman, head of product at Meta Superintelligence Labs, said 1Password gives Meta's agent safe access to services at the user's direction. Jean-Denis Greze, co-founder and CEO of Town, said Townies can securely reach the tools people use while their credentials stay protected.
“With 1Password, Grok Bot can sign in when it needs to without asking people to share or paste their passwords.”— Tido Carriero, Head of Product, SpaceXSI
Beyond the consumer agents, 1Password lists Browserbase, Browser Use, Dana, Instinct, KERNEL, Perplexity, Ramp, Tomo and Wajo as partners, and says more integrations are coming. The announcement does not include pricing. John Torrey, 1Password's chief business officer, described the partner list as the industry working together to build agents people trust.
The launch builds on a Mac integration for Claude that 1Password shipped in July, which filled credentials into the browser without returning the plaintext password to the model and granted access one task at a time. At the time, 1Password CTO Nancy Wang said protecting the credential does not limit what an agent does after it signs in, which leaves prompt injection as a separate risk.
The launch coincided with a report that U.S. banks have begun detecting agentic activity in customer accounts, and that most banks cannot tell whether that activity reflects a request from the account holder. A Chase spokesperson said customers should not give their usernames and passwords to anyone.
1Password says its enterprise vault already protects more than 1.6 billion credentials and secrets and is used by over 200,000 businesses.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
