Skip to content
Cyberttak.
Autonomous agents· 3 min read

OpenAI's dots, always-on agents on GPT-6 Astra, connect to more than 4,000 apps

The agents launched at DevDay on September 29, 2026 for ChatGPT Pro and Business Premium users in eligible markets, with the first dot included.

By CyberttakPublished · 07:32

In brief

OpenAI dots plug into 4,000+ apps

OpenAI launched dots at DevDay on September 29, 2026. Each agent has its own cloud computer and works around the clock. It starts with Pro and Business Premium users in eligible markets.

The OpenAI logo viewed through a magnifying glass
The OpenAI logo viewed through a magnifying glassCredit: Photo: Jernej Furman from Slovenia / Wikimedia Commons · CC BY 2.0 (cropped)

OpenAI has launched dots, personal AI agents that work toward a user's goals around the clock. The company announced them at its DevDay conference in San Francisco on September 29, 2026. We know this from OpenAI's own launch post and from reports by CBS News and Al Jazeera.

What dots are

An agent is an AI that does tasks for you instead of only answering questions. OpenAI says each dot has its own cloud computer and browser. It runs on GPT-6 Astra, a model OpenAI released earlier in September, according to CBS News.

Dots can connect to more than 4,000 apps through plugins, including Slack and Teams. CBS News lists tasks such as scheduling meetings, booking travel, budgeting, debugging systems and writing code.

Who gets it, and at what price

Access starts with ChatGPT Pro and Business Premium users in "eligible markets", and the first dot is included. CBS News reports that Pro starts at $100 a month and Business Premium at $20 a month. Other outlets give different figures or details on the plans, so check OpenAI's pricing page before paying.

Sources also differ on enterprise access. Times of AI and Lowyat.NET describe it differently, so the exact terms for large companies are not settled here.

Working when you are away

OpenAI describes a read-only mode for when the user is absent. Al Jazeera reports that it stops agents from controlling the user's browser or computer in that time. OpenAI calls this one of several extra safety and privacy safeguards.

Imagine you leave work on Friday and ask a dot to prepare notes on your Monday meeting. In that mode it could read and gather material, but it could not take actions on your behalf.

From coding tools to dots

The story of where dots came from is less tidy. Mezha and Jingletree report that the team spun ChatGPT Work out of agentic coding tools. The product lead, Embiricos, said much of the functionality is also possible with Codex, which is a narrower claim than saying dots simply grew from coding tools. Embiricos previously co-founded Multi, which OpenAI acquired in June 2024, according to Cryptobriefing and NewsBytes.

Context

CEO Sam Altman said people can delegate big pieces of work as they would to a chief of staff. He also said that in future people will be able to talk to agents by phone and text. Meta released its Muse agent earlier in September, and Google unveiled Gemini Spark in May.

Safety worries sit in the background. The day before the launch, OpenAI said it would not release GPT-6.1 Astra after in-house testing found it did not stay within scope and authorization. Al Jazeera says OpenAI is under scrutiny over hacking incidents involving its agents.

Would you hand a dot your calendar and your budget? The sources show the launch and the safeguards OpenAI describes. They do not show independent tests of how reliably dots behave, and they give no date for phone and text access.

Sources