OpenAI's headline act at DevDay wasn't a new model. At the keynote on Tuesday, the company introduced dots: "remarkably capable, always-on agents built to handle everything," powered by GPT-6 Astra.

Introducing dots, powered by GPT-6 Astra.

Remarkably capable, always-on agents built to handle everything. pic.twitter.com/inY5BZj3Kz

— OpenAI (@OpenAI) September 29, 2026

On Monday we wrote about leaked strings for an always-on Pro assistant called "o", and on Tuesday afternoon a rumor put the name as Dots. The leaks were right about the product but not the name. It launched as "dots," lowercase. In hindsight, OpenAI's teaser, "10am PT, on the dot," was a hint.

According to OpenAI, each dot gets its own cloud computer and browser, connects to over 4,000 apps through ChatGPT's plugins, and learns from your feedback. The idea is to hand it an ongoing responsibility, not a single prompt. In one of OpenAI's examples, a dot watches customer feedback, builds and tests small fixes, and brings back pull requests for you to review. An early tester's dot noticed he'd forgotten to invoice a publication, prepared the invoice and sent it once he approved.

You can message or call your dot in ChatGPT on the web, desktop and mobile, and message it in Slack and Teams. Texting is "coming soon." You start with one primary dot that you can name, and teams of dots will come later.

Control is the obvious worry. Dots start with built-in rules for when to act and when to ask. Custom Rules let you allow specific actions, require approval or block them, and an Auto-review system checks anything that could affect your accounts or share information. When you're not around, your dot does "proactive research" with read-only tools that can't send messages or change anything. Its safety post says saved passwords are filled in without being exposed to the model, and connecting your own computer is optional.

That's a lot of trust to ask for in a month when OpenAI has disclosed several incidents involving its own agents. OpenAI's advice: "Dots can still make mistakes, so always review consequential work."

Dots are rolling out to Pro and Business Premium users in eligible markets. Enterprise, Edu and Healthcare workspaces get a beta that admins must switch on. The release notes add two catches: you must be 18 or older, and Pro access excludes the European Economic Area, Switzerland and the UK at launch. OpenAI's Tibo says the $100 Pro plan is included. You'll need to set up your dot on the desktop app or desktop web first.

The usage rules aren't fully settled, and OpenAI describes them two ways. The announcement says "conversations with your dot don't count toward your ChatGPT usage limits," but tasks it starts in Codex or ChatGPT Work "count toward your usage limits as usual," with extended limits for the first month. The release notes say: "For the next month, dots usage won't count toward eligible Pro, Business, and Enterprise users' plan allowances." OpenAI says it will share usage terms for each plan after that.

For companies, OpenAI is previewing "specialist dots" with their own identity and credentials, starting with enterprise pilots, and plans to manage them through Microsoft's Agent 365.

Dots join a crowded field that includes Meta's Muse and Microsoft's Autopilot. OpenAI's pitch is its strongest model, running around the clock. The real test is whether people are happy to give it their inbox.