Key facts
- 1Password for Agents uses a unique digital key to confirm an agent's identity.
- Users receive an alert to confirm login requests from agents.
- The service supports agents from Anthropic, OpenAI, Meta, and Town.
Password manager 1Password has introduced 1Password for Agents, a new service designed to address concerns about AI agents accessing user accounts. The tool enables users to grant AI agents permission to log into their personal accounts for specific tasks without exposing their actual passwords or credentials. When an agent attempts to log in, 1Password verifies the agent's digital identity via a unique key and sends an alert to the user for confirmation. Upon user approval, 1Password facilitates the login, allowing the agent to complete its assigned task. The service supports a range of AI agents, including those from Anthropic, OpenAI, Meta, and Town, with plans to expand compatibility. The launch coincides with reports of U.S. banks detecting increased agentic activity within customer accounts, highlighting the growing need for secure methods of AI interaction.