OpenAI's new always-on ChatGPT dots exclude Pro users in the EEA, Switzerland and the UK
OpenAI has launched dots, always-on agents inside ChatGPT that keep making progress between conversations and run on GPT-6 Astra, and Pro access to them "excludes the European Economic Area, Switzerland, and the UK at launch". The model and the exclusion are both stated in OpenAI's ChatGPT release notes for September 29, which also set an age floor of 18.
ChatGPT in a desktop browser, with Library, Projects, Scheduled and Plugins in the sidebar and a conversation about cooking dinner from what is in the fridge. Image: OpenAI Image: OpenAI
Business Premium subscribers are the exception to the geography. OpenAI's help article says dots reach them "across all supported ChatGPT regions", while Enterprise workspaces get a beta that is off until an admin enables it. Dots are one of the products OpenAI announced at its DevDay event on September 29, and it says the rollout is gradual, so access "may take several days to reach your account".
Each dot gets its own computer, and may reach yours
A dot has its own cloud computer and its own browser, and OpenAI's help article says access to a local computer is optional and "starts turned off". Connecting one takes the ChatGPT desktop app on that machine and a confirmed "Allow access", which "Revoke access" undoes.
With local access on, a dot can create Work or Codex tasks, use local skills and use the local browser when its cloud browser is blocked. Creating a dot is desktop-only, in the ChatGPT desktop app or on desktop web, and dots are not supported on mobile web at all.
Proactive research runs on read-only tools
A dot works across the apps a ChatGPT account connects, and it does not wait to be asked. It "can also review that information proactively and form memories from it, even when you haven't asked a specific question about it", OpenAI says. It calls that background work proactive research, and its announcement puts it on tools "restricted to be read-only", which cannot send messages, change app content or control a browser or computer.
Custom rules govern everything else, with four behaviours: "Take action without asking", "Take action if pre-approved", "Ask before taking action" and "Hand off to you". OpenAI defines the second one narrowly: "'Pre-approved' means you explicitly requested the action in your prompt."
Texting is a US Pro beta on a third-party number
Texting "uses a third-party provider and is offered as a limited beta", restricted to Pro users in the US and unavailable in Business or Enterprise workspaces. Replying STOP ends outgoing texts, message and data rates may apply, and OpenAI states plainly that a dot "cannot initiate calls to you at launch".
Dots usage will not count towards eligible Pro, Business and Enterprise plan allowances "for the next month", after which OpenAI says it will share usage terms for each plan. Deleting a dot is the only way to clear what it has picked up, because disconnecting an app "does not delete information your dot has already obtained from it".
How well any of it works is not established. OpenAI has published no benchmark, no success rate and no independent evaluation for dots, and its own caution is that a dot "can make mistakes, including when following your rules".
