Briefing · 30 September 2026 · Australia/Brisbane

Dots

OpenAI’s always-on agents inside ChatGPT, announced on a page dated 29 September 2026. This page only restates what those public documents say.

What it is

On Introducing dots (dated 29 September 2026), OpenAI describes dots as “remarkably capable, always-on agents built to handle everything.” They are powered by GPT-6 Astra, have their own cloud computer, learn from feedback over time, and can work toward a person’s goals around the clock. Through ChatGPT plugins they can connect to over 4,000 apps. A person reaches a dot in ChatGPT (desktop, web, and mobile), Slack, or Microsoft Teams, including by voice call. OpenAI says texting is coming soon on that announcement; the Help Center, updated the same rollout, separately describes a limited texting beta (below).

A dot can keep several projects going, use its own browser and computer, and optionally be allowed to use the person’s laptop. While the person is not actively working with it, “proactive research” can look through already connected apps with read-only tools: those tools cannot send messages, change app content, or control a browser or computer. Follow-up actions go through the usual permissions. Custom Rules can allow, require approval for, or block actions, but built-in safety checks still apply. Sensitive steps such as changing a password stay with the person. Auto-review checks actions that could affect accounts or share information. Each dot runs on its own cloud computer; the person’s computer stays separate unless they connect it. For supported website sign-in, saved passwords are not shown to the model.

The rollout stated on the announcement and on Getting started with your dot is to ChatGPT Pro and Business Premium in “eligible markets,” plus an Enterprise beta (including Edu and Healthcare) that a workspace admin must turn on. The first dot is included in Pro or Business Premium at no extra charge. Conversations with the dot do not count toward ChatGPT usage limits. Tasks the dot starts in Codex or ChatGPT Work still count toward those products’ limits. The plan includes an allowance for deeper work, with extended limits for the first month after launch. OpenAI says that later people will be able to add more dots and scale each dot’s speed or monthly amount of work. Setup is in the ChatGPT desktop app or a desktop browser, including Windows. A dot cannot be created on mobile or on mobile web; after setup, messaging in the mobile app is available where that access has rolled out. The Help Center says features announced at DevDay are rolling out gradually and may not be on an account yet.

Dots are not a separate app from ChatGPT. Plugin permissions are shared with ChatGPT, ChatGPT Work, and Codex. A dot can create Codex cloud tasks if the environment already exists, and, if local computer access is turned on, can create Work or Codex tasks on that computer. Enterprise model controls do not apply to dots. “Specialist dots” are a separate idea: company-provisioned agents with their own identity for defined jobs, starting as focused enterprise pilots. OpenAI says it is working with Microsoft so specialist dots can be managed through Agent 365. That integration is described as a goal, not as a finished public product.

OpenAI, Introducing dots, 29 September 2026: “Dots are remarkably capable, always-on agents built to handle everything.”

Documents

No dots-specific page was found on developers.openai.com, and no dots GitHub repository was found. The Agents API announced around the same DevDay is a separate developer product; the dots announcement does not present it as the way to create a ChatGPT dot.

Videos

Official OpenAI uploads and embeds first, then independent explainers posted 29 September 2026 Pacific time (30 September 2026 in Brisbane). Times below are Australia/Brisbane. Lengths are from YouTube. These were not watched end to end; each “why” line is from the title, publisher, and length.

Official

Independent explainers

What is still unknown