OpenAI rolls out always-on dot agents in ChatGPT
OpenAI is rolling out dots, always-on agents in ChatGPT that can research, create content, write and test code, and manage recurring tasks between conversations. Each dot has its own cloud computer, draws on ChatGPT memory, and can build memories from connected apps. Access to a user’s local computer is optional and disabled by default.
Availability includes Pro users aged 18 and older outside the European Economic Area, Switzerland, and the UK, plus eligible Business Premium accounts in supported regions. Eligible Enterprise, Edu, and Healthcare workspaces also have access, subject to administrative controls and residency restrictions. Personal and Business Premium access is rolling out gradually. Enterprise dots have no fee through January 13, 2027. Billing begins January 14, 2027, with delegated ChatGPT Work or Codex tasks billed separately.
Users can set action rules, require approvals, review activity, and pause their dot. Connected information can be reviewed proactively, even without a new request. Disconnecting an app does not erase information already obtained, and pausing a dot may leave separate delegated or scheduled tasks running. Deleting a dot starts deletion of its own saved memories but retains separate Work or Codex chats.
Texting is in beta for limited Pro users in the US, Canada, and Mexico through iMessage or RCS. Messages pass through provider Linq and are not end-to-end encrypted between users and their dots. OpenAI stores the texts in ChatGPT conversations; Linq does not retain message contents after delivery.