ChatGPT Dots Put Always-On Agents to Work Between Chats

OpenAI introduced ChatGPT dots at DevDay on September 29, 2026. A dot is an always-on agent that can keep working between conversations, using its own cloud computer, connected apps, and saved context to handle ongoing tasks and bring results back for review. (OpenAI's DevDay recap; OpenAI Help Center; Tibo on X)

Key Takeaways

• A dot can continue work between conversations, run scheduled tasks, and ask you to make decisions that need your judgment. (OpenAI Help Center)
• Each dot has a cloud computer and can use apps you choose to connect. Access to your local computer is optional and starts turned off. (OpenAI Help Center)
• OpenAI includes a first dot at no extra cost with Pro and Business Premium; each plan also has an allowance for deeper work. (OpenAI Help Center)
• Pro availability excludes the European Economic Area, Switzerland, and the United Kingdom. Business Premium is available in supported ChatGPT regions; Enterprise, Edu, and Healthcare users can try an admin-enabled beta. (OpenAI Help Center; OpenAI)
• Dot creation currently requires the ChatGPT desktop app or desktop web. Access is rolling out gradually, and some features may not have reached every account. (OpenAI Help Center)

How Dots Handle Ongoing Work

A dot uses a goal and the context you provide to work across connected apps, remember information relevant to ongoing work, and keep making progress between messages. OpenAI says dots can research in the background, review connected information proactively, run reminders or recurring tasks, and return to you when they need a decision that requires your judgment. (OpenAI Help Center)

Each dot has its own cloud computer. Connecting your own computer is optional and starts disabled; after you grant access through the ChatGPT desktop app, the dot can work with local files and use local skills. OpenAI also documents rules that let you decide which actions a dot can take on its own, take only when pre-approved, or hand back to you. (OpenAI Help Center)

Availability, Cost, and Platform Limits

OpenAI's current help guidance gives dots a staged rollout rather than one universal launch date. The first dot is included at no extra cost for Pro and Business Premium users, but access depends on plan, region, and account rollout. Enterprise, Edu, and Healthcare users can try a beta when an administrator enables it; the beta is off by default. (OpenAI Help Center; OpenAI's DevDay recap)

OpenAI says the plan includes an allowance for deeper work, with extended limits for the first month after launch. It is also rolling out dots gradually, so an eligible user may need to wait several days before access appears. (OpenAI Help Center)

Practical Guardrails for a First Dot

Start with one bounded task that benefits from continuity, such as a recurring check or research task, and connect only the apps it needs. Set explicit rules for actions involving sharing, purchases, or access, and choose to be asked when an action needs your approval. OpenAI warns that dots can make mistakes, including while following custom rules, so review important work before relying on it. (OpenAI Help Center)

You can inspect scheduled and completed work in the dot's profile, pause a dot when you do not want it to continue, and revoke local-computer access. Those controls make it practical to test a dot on a low-risk workflow before delegating more consequential work. (OpenAI Help Center)

Sources

• OpenAI, "DevDay 2026 Recap," September 29, 2026: https://openai.com/index/devday-2026-recap/
• OpenAI Help Center, "Getting started with your dot": https://help.openai.com/en/articles/20001530-getting-started-with-your-dot
• OpenAI, "Meet your dot," ChatGPT Release Notes: https://help.openai.com/en/articles/6825453-chatgpt
• Tibo on X, "Announcing dots," September 29, 2026: https://x.com/thsottiaux/status/2104981170685616361
• OpenAI on X, "Introducing dots, powered by GPT-6 Astra," September 29, 2026: https://x.com/OpenAI/status/2104984504133918973