Quick Answer
OpenAI Dots are always-on AI agents that work toward user goals around the clock on separate cloud computers. OpenAI launched Dots on September 29, 2026, for eligible ChatGPT Pro and Business Premium users, with administrator-enabled beta access for some managed plans. Set up a Dot from a desktop first, review every connected app, and keep financial or account changes under your direct control.
Key Takeaways
- OpenAI Dots use GPT-6 Astra and operate from separate cloud computers.
- ChatGPT Pro and Business Premium users receive their first Dot at no additional cost.
- Dots can connect to more than 4,000 apps through plugins, according to OpenAI.
- Background research is limited to read-only tools and cannot change connected-app content.
- Purchases, password changes, and financial transfers have approval or user handoff limits.
What are OpenAI Dots?
OpenAI Dots are always-on AI agents designed to work toward a user’s goals around the clock. OpenAI announced the product on September 29, 2026, and says each Dot runs on its own cloud computer using GPT-6 Astra. OpenAI’s Dots announcement describes the service as an agent system that can continue work outside of an individual chat session.
OpenAI Dots differ from a standard chatbot interaction because the agent has a dedicated cloud environment rather than relying only on the device in front of the user. That separation matters because a Dot can use its own browser and cloud computer while the user’s local computer remains separate unless the user explicitly connects it. The practical benefit is that a task can continue without occupying a personal laptop, but users still need to decide which accounts and services the agent can access.
OpenAI Dots also reflect a broader move toward AI systems that can carry out longer-running work rather than answer one prompt at a time. Users considering persistent AI tools should review ChatGPT privacy settings before connecting services that contain personal, work, or financial information.
Who can use OpenAI Dots at launch?
OpenAI Dots initially roll out in eligible markets to ChatGPT Pro and Business Premium users. OpenAI says Enterprise, Edu, and Healthcare customers can access a beta version after an administrator enables the feature, which means managed-account users cannot necessarily activate Dots individually.
OpenAI includes a user’s first Dot at no additional cost with a Pro or Business Premium plan. The company also says those plans include extended deeper-work limits for the first month. The offer is tied to the eligible plan rather than presented as a separate standalone product price, so users should confirm current plan eligibility in ChatGPT before depending on Dots for routine work.
OpenAI’s rollout language also matters because access depends on market eligibility and account type. A user who does not see Dots in ChatGPT should not assume that an account problem exists, particularly when the feature is still rolling out or an organization controls product access.
How do you set up an OpenAI Dot?
OpenAI Dots require initial setup through the ChatGPT desktop app or a desktop browser. OpenAI says users can message a Dot through the ChatGPT mobile app after setup, which makes mobile access available for follow-up communication rather than first-time configuration.
- Open ChatGPT in the desktop app or a desktop browser.
- Create the OpenAI Dot available to your eligible account.
- Review the goal you want the Dot to pursue before connecting any plugin.
- Connect only the apps and services needed for that goal.
- Use the ChatGPT mobile app to message the Dot after desktop setup is complete.
OpenAI Dot setup is a security decision as well as a convenience step because plugins can extend the agent into connected services. Start with the smallest practical set of connections, especially for work accounts. Stop and ask an organization administrator before connecting a managed Enterprise, Edu, or Healthcare account if company policy or account permissions are unclear.
What apps and communication channels can OpenAI Dots use?
OpenAI Dots can connect through plugins to more than 4,000 apps, according to OpenAI. The company says users can contact a Dot through ChatGPT, Slack, and Microsoft Teams, while text messaging support is listed as coming soon. Those channels give Dots a way to receive direction where users already communicate, but the available connections will depend on the apps a user chooses to authorize.
OpenAI’s plugin model makes Dots potentially useful for work that spans several services, such as collecting information from approved sources and returning results through a communication platform. At the same time, more connections create a broader access footprint. The most sensible approach is to connect one service at a time, verify what the Dot can see, and remove access that no longer supports an active goal.
OpenAI Dots enter a consumer-agent market that includes other products intended to carry out more persistent tasks. Meta’s own personal agent effort has taken a different form through Meta’s pocket AI device, which shows that companies are exploring both cloud-based and hardware-based approaches to personal assistance.
How does OpenAI limit what Dots can do in the background?
OpenAI Dots restrict proactive background research to read-only tools. OpenAI says that restriction means a Dot cannot send messages, change content in connected apps, or directly control a browser or computer while conducting background research. The limitation is important because it separates information gathering from actions that could affect another person, alter records, or create an external commitment.
OpenAI’s safety material describes the controls around GPT-6 Astra and Dots, including the limits applied to agent actions. OpenAI’s GPT-6 Astra change log explains the product’s safety approach. The stated restrictions reduce the scope of unattended work, but users should still treat any connected-app access as sensitive because the Dot can work with information made available through approved tools.
| OpenAI Dot activity | OpenAI’s stated limit | Practical user response |
|---|---|---|
| Background research | Restricted to read-only tools | Connect sources that are appropriate for information access. |
| Sending messages | Not allowed during proactive background research | Review messages yourself before they are sent. |
| Changing app content | Not allowed during proactive background research | Keep editing and publishing decisions under direct user control. |
| Browser or computer control | Not allowed during proactive background research | Do not assume unattended research can perform interactive tasks. |
Which OpenAI Dot actions require user approval?
OpenAI Dots require approval for purchases that use cards already saved at a merchant. OpenAI also says password changes and transfers between financial accounts must be handed back to the user. Those boundaries matter because purchases, credential changes, and money movement can create immediate financial or account-security consequences that are difficult to reverse.
OpenAI’s approach places a human approval step at the point where an agent could spend money or alter account control. Users should still review the merchant, the purchase amount, and the destination account before approving an action. An approval prompt is not a substitute for checking whether a request is legitimate, necessary, and consistent with the user’s original goal.
OpenAI Dot users should keep authentication details, financial accounts, and highly sensitive documents out of unnecessary plugin connections. Recent concerns around AI agent data exposure also show why users should limit the information an autonomous system can access, even when the product includes stated safeguards.
How should users decide whether OpenAI Dots are useful?
OpenAI Dots are most useful for users who need an agent to continue a defined goal over time and can describe the boundaries clearly. The product’s separate cloud computer, multi-app plugin support, and communication options make it suited to longer-running research and coordination tasks, while the read-only background limit reduces the chance that unattended work will change external systems.
OpenAI Dots are less appropriate for users who need fully autonomous purchasing, account management, financial transfers, or unsupervised browser control. OpenAI requires user approval or handoff for several of those higher-risk actions, which means the product is not designed to replace judgment at the point where money, credentials, or account ownership are involved.
OpenAI positioned Dots as its main consumer-focused DevDay announcement, while Axios described the launch as competition for Meta’s Muse personal agent. Axios’s DevDay coverage provides that market context. For most users, the practical test is simple: start with a narrow task, connect minimal data, and expand access only after the Dot’s behavior matches the intended workflow.
FAQ
Are OpenAI Dots available to all ChatGPT users?
OpenAI Dots are initially available in eligible markets to ChatGPT Pro and Business Premium users. Enterprise, Edu, and Healthcare users need an administrator to enable beta access.
Does an OpenAI Dot use my personal computer?
OpenAI Dots use their own cloud computers and browsers by default. OpenAI says a user’s local computer remains separate unless the user explicitly connects it.
Can OpenAI Dots send messages while working in the background?
OpenAI Dots cannot send messages during proactive background research. OpenAI restricts that research mode to read-only tools, which also prevents content changes and direct browser or computer control.
Can an OpenAI Dot make purchases?
OpenAI Dots can use cards already saved at a merchant only after user approval. OpenAI requires the user to approve the purchase rather than allowing the agent to complete it independently.
Can OpenAI Dots change passwords or transfer money?
OpenAI Dots must hand password changes and transfers between financial accounts back to the user. Those actions remain under direct user control because they can affect account security and financial ownership.