OpenAI dots, explained: how an always-on AI agent keeps working after you close the chat

On 29 September 2026, at its DevDay event, OpenAI announced dots. It calls them "always-on agents".

An agent is an AI model that does more than answer. It takes actions one after another, such as opening a web page, editing a file or drafting a message, to reach a goal you gave it.

OpenAI's launch page says dots are "Powered by GPT-6 Astra" and that through plugins they "can readily connect to over 4,000 apps". A plugin, also called a connected app, is a link between ChatGPT and another service, such as your calendar or your code host. It lets the agent read or change things there with your permission.

You can talk to your dot in ChatGPT on desktop, web and mobile, and in Slack and Microsoft Teams. Texting is a limited beta for some Pro users in the US.

OpenAI also showed a preview of "specialist dots" for companies, each with its own identity and access to company systems, and says it is working with Microsoft to manage them through Agent 365.

In an ordinary chat, the model answers your message and then waits for the next one. A dot does not wait.

Why does that matter? Some work does not fit in one sitting. Checking your calendar every morning, watching customer feedback for repeated bug reports, or following a news story for a week all need something that keeps running when you are not looking.

A dot can start work without a new message from you in two ways. It runs reminders and recurring tasks you scheduled. And it does what OpenAI calls "proactive research": it reads the apps you connected, looks for ways to help and saves private notes.

OpenAI says these research tasks use read-only tools, enforced in code, that "cannot directly send messages to other people, change content in connected apps, or control a browser or desktop". Anything the dot then wants to do goes through the normal approval rules.

You can still see and stop the work. The dot's profile lists tasks under In progress, Scheduled and Completed. You can open its computer to view it and interact with it, you can Pause it, and Reset deletes the dot with its conversations, saved memories and scheduled tasks.

Memory needs a careful read. A dot receives memories from ChatGPT and creates its own, including from your connected apps. According to OpenAI's privacy FAQ, you cannot currently view or delete individual dot memories, and disconnecting an app does not delete what the dot already learned from it. Deleting the dot is the way to delete its context.

We did not use a dot. Dots are a hosted ChatGPT feature for paid plans, with no local version, so there was nothing we could run or measure on our own machine. That is why this explainer has no hands-on section and shows no screens.

Every statement about dots comes from four OpenAI pages and one recap on OpenAI's developer forum, all read on 4 October 2026 and linked below. OpenAI's site refused a plain download with a 403 error, so we read the pages in an ordinary Chrome browser. The model name, GPT-6 Astra, is stated on OpenAI's own launch page and help page.

OpenAI did not announce a public API for dots in the pages we read. Its developer recap lists a separate product, the Agents API, in public beta, with hosted execution, memory and tools. That is a different product for developers, and we did not test it.