OpenAI launches Dots personal AI agent to compete with Meta’s Muse

OpenAI has launched Dots, an always-on AI agent designed to compete directly with Meta’s Muse, intensifying a race to control the personal AI assistant category that both companies now see as central to consumer AI. The rollout signals how quickly the two firms are converging on nearly identical products, raising questions about safety guardrails that Meta is already facing scrutiny over.

• OpenAI announced Dots during its DevDay keynote on Tuesday (September 29), powered by its GPT-6 Astra model.
• Dots can access a cloud computer and more than 4,000 connected apps, plus Microsoft Teams and Slack.
• The feature is rolling out now to ChatGPT Pro and Business Premium subscribers, with only one Dot allowed per user initially.

• 4,000+ apps a single Dot can connect to and operate across
• 1 Dot allowed per user at launch versus planned multi-agent future
• Sep 29 DevDay 2026 date Dots was unveiled to the public

OpenAI unveiled Dots at its DevDay 2026 keynote on Tuesday (September 29), positioning the product as its answer to Meta’s Muse. Dots are described by OpenAI as always-on assistants that can “do nearly anything” across a user’s connected apps while working in the background.

The timing is notable. Meta launched Muse just weeks earlier, and the assistant has surged in downloads even as safety concerns mount, including a user report that Muse disclosed their home address to a Facebook Marketplace buyer without permission. OpenAI’s response arrives fast enough to suggest the company views agentic personal assistants as an immediate competitive front rather than a longer-term roadmap item. The story was first reported by The Verge.

GPT-6 Astra gives Dots a cloud computer and 4,000-app reach

Dots run on OpenAI’s GPT-6 Astra model and operate from their own cloud computer, giving each assistant a persistent environment to browse the web and control more than 4,000 supported applications. Users interact through a text-message-style interface similar to Muse’s, or by placing a voice call to a Dot from ChatGPT on web, desktop, or mobile.

Dots also connect to Microsoft Teams and Slack, carrying context between sessions across devices and apps. OpenAI said text-message access is coming soon as well.

OpenAI’s examples illustrate the intended use cases: a developer could direct a Dot to build and test app updates from customer feedback and receive a video of the changes, while a content creator could have a Dot review an interview transcript, flag clip-worthy moments, draft show notes, and write social posts. Users can monitor a Dot’s progress by viewing its cloud computer directly, and the assistant will message with updates or questions as it works.

OpenAI limits users to one Dot while building in permission rules

For now, OpenAI is capping each account at a single Dot, though the company has said it plans to let users deploy multiple agents simultaneously. Dots come with what OpenAI calls “built-in rules for when to act independently,” alongside custom controls that let users block specific actions or require the assistant to ask permission first.

An “auto-review” feature checks whether a planned action complies with those custom rules and broader safety requirements. OpenAI says this system also determines which tasks, such as changing a password, get handed back to the user rather than completed autonomously.

That framing stands in contrast to the safety criticism Meta’s Muse has drawn, though OpenAI has not commented on Muse’s specific incidents and Meta’s response to that reporting is not addressed in the available coverage.

ChatGPT Space and enterprise “specialist” Dots are next

OpenAI is also introducing ChatGPT Space, a collaborative workspace where users and colleagues can interact with a shared Dot. Separately, the company is testing “specialist” Dots that let organizations create agents assigned to specific internal roles.

OpenAI said conversations with a Dot will not count against standard ChatGPT usage limits.

In the future, you’ll be able to add more dots, and scale the output of each dot by either increasing its speed or the total amount of work it can take on per month.

OpenAI

The BlockWest read. The real contest here is not features but trust infrastructure: whichever company can prove its auto-review and permission systems actually prevent Muse-style leaks will win enterprise adoption first. Business Premium customers, not consumers, are the near-term test case, since Slack and Teams integration puts Dots inside corporate data flows where a single unauthorized action carries real liability.

OpenAI has not set a date for expanding Dots beyond a single agent per account or for opening the specialist-agent feature beyond its current test group, leaving enterprise customers to watch for the next DevDay update on both fronts.

BlockWest is a news publication. Nothing here is investment advice. Read our disclaimer and editorial policy.