1Password is introducing Unified Access Pro, a platform that manages credentials and access for people, AI agents, and machine identities simultaneously. With device-level visibility, centralized storage, and just-in-time credential delivery, the company aims to help organizations safely adopt AI tools.
One problem Unified Access Pro aims to address: exposed SSH keys, plaintext .env files, and long-lived API tokens rarely appear in traditional SaaS logs or federated identity systems. Yet they can grant direct access to production environments. As more employees experiment with AI agents, the volume of unmanaged credentials in circulation increases.
Unified Access Pro brings visibility down to the device level. It detects AI tools and local agent activity on devices and in browsers, identifies exposed credentials, and maps findings back to users and devices. Credentials are then centralized in a single secure vault, with consistent policies for people, agents, and machine identities. Notable is the shift from “always-on” access to just-in-time credential delivery: an agent receives credentials only when they are actually needed.
Partners ranging from Anthropic to Cursor and GitHub
The launch is accompanied by a broad network of partners. Anthropic and OpenAI are collaborating with 1Password to deploy vault items within agentic browser flows and developer IDEs. Cursor, GitHub, and Vercel are integrating with the platform for CI/CD pipelines and cloud environments. For MCP gateways, 1Password connects to Natoma and Runlayer, while AI browsers such as Perplexity and Browserbase can retrieve credentials just-in-time with least-privilege controls.
Unified Access Pro is now generally available as a separate tier for security teams.