1Password has released 1Password for Claude, a new feature on its browser extension that lets Claude use your credentials while keeping private data from reaching the model, its memory, or Anthropic's servers. As 1Password explains, users can authorize Claude to complete real‑world tasks—booking travel, managing accounts—securely with credentials injected directly to the target system on their behalf.

Security has emerged as a major concern when agents perform tasks on your behalf. Sharing credentials directly with an agent exposes them to the model, its memory, and the systems behind it, which can pose a risk if the model is compromised. To address this, 1Password built a zero‑exposure security framework that allows agents to use stored credentials from your vault without them reaching the model.

Access is granted per session, limited to specific, approved items and not carried over. Passwords and one‑time MFA codes are injected through a secure channel managed by 1Password, keeping them outside the model. The service also enables authenticated multi‑site sessions to complete multi‑step workflows without repeatedly prompting the user, and 1Password locks down when an AI agent takes control, limiting access to only the credentials explicitly granted for the current task.

Alongside Claude integration, 1Password introduced an Agentic Mode that safeguards credentials against active agents. When detected, the extension locks down so that the agent can only find passwords explicitly granted for the task, and users see the security feature active in the browser extension. 1Password for Claude is now available to zap users on business, family, and individual plans, requiring the 1Password desktop app and browser extensions, plus the Claude desktop app and browser extensions. Future support for payment cards and identity details will follow.