OpenAI launched Dots at its DevDay event on September 29, 2026, introducing background agents powered by the GPT-6 Astra model that retain context across interactions.

Each Dot runs on an isolated cloud computer and can connect to more than 4,000 applications, including ChatGPT, Slack and Microsoft Teams.

With a user’s permission, Dots can retrieve verification codes from email and enter them during sign-up or login, potentially carrying multi-step tasks forward without human intervention.

Access is limited to ChatGPT Pro and Business Premium subscribers, who get one Dot with qualifying subscriptions; an Enterprise beta is also available.

OpenAI says safeguards include an Auto-review process, user-defined Custom Rules and limits on high-risk actions such as confirming transactions or changing sensitive data.

Early user feedback suggests Dots may still pause before entering a two-factor authentication code, even when they can read it, reflecting a cautious approach to security.