- OpenAI Dots are always-on AI agents powered by GPT-6 Astra, launched at DevDay on Sep 29, 2026
- Pro workflows depend on each Dot's dedicated cloud computer, browser, and 4,000+ app connections
- Autonomy settings control how independently a Dot acts between conversations
- Permissions cover connected apps, computer access, and workspace-level admin controls
- Scheduled tasks turn a Dot into a persistent assistant for recurring work
What OpenAI Dots Pro Users Get
OpenAI Dots, announced at DevDay on September 29, 2026, are persistent AI assistants built on GPT-6 Astra. Unlike a normal ChatGPT conversation, each Dot keeps working toward assigned goals across multiple sessions, using its own cloud computer, its own browser, and connected apps such as Slack and email. For pro-level users, the difference is continuity: context, files, and task history persist, so long-running work never starts from zero.
Astra is positioned by OpenAI as its flagship model for complex reasoning, coding, computer use, and research, which is what makes multi-step agentic work practical. OpenAI's official safety documentation has also added a dedicated Dots section covering how these agents are governed.
| Capability | What It Does | Pro Value |
|---|---|---|
| Always-on agent | Works between conversations toward a goal | Long-running projects without re-prompting |
| Cloud computer | Dedicated remote machine per Dot | Browser tasks run without your device |
| App connections | 4,000+ connected tools and channels | Cross-service workflows |
| GPT-6 Astra | Flagship reasoning and computer-use model | Complex multi-step execution |
| Follow-up channels | ChatGPT, email, Slack, SMS | Proactive updates where you work |
Frame every Dot around an ongoing outcome, not a single prompt. A goal like "keep my research briefing current" outperforms a one-off request, because the Dot is designed for persistent objectives.
Setup: From Creation to First Task
Configuration quality decides how useful a Dot becomes. Start from ChatGPT on desktop web or the supported desktop experience, then follow this sequence:
Create the Dot
Open the Dots experience in ChatGPT and start a new Dot. Name it after the job it owns, such as "Research Desk" or "Inbox Manager."
Define the Goal
Write a clear ongoing outcome. A focused goal gives the Dot stable direction across many work sessions.
Set Working Instructions
Describe priorities, boundaries, and how independently the Dot should operate. Tighter boundaries mean more approval checkpoints.
Attach Context and Files
Add project documents and reference material so the Dot has the background it needs from day one.
Connect Apps
Link supported services such as Slack and email. Each connection uses only the permissions you or your workspace grant.
Assign the First Task and Review
Give a concrete task tied to the goal, then monitor progress from the activity dashboard, where active and completed work is logged.
Advanced configuration is centered on desktop web and the desktop app. Mobile access varies by account and rollout stage, so do your initial setup on desktop.
Pro Automation Patterns
The strongest pro use cases are workflows where context must persist. Four patterns cover most needs:
Scheduled Recurring Work
- Repeating checks on a set schedule
- Recurring summaries and digests
- Runs without manual prompts
Recurring Monitoring
- Watch information that changes over time
- Surface meaningful updates only
- Ideal for tracking topics and sources
Connected-App Workflows
- Combine actions across Slack, email, files
- Cross-service multi-step execution
- Boundaries set by granted permissions
Ongoing Research
- Collect, organize, and update findings
- Continues across multiple sessions
- Evolves as the topic changes
Calendar and Follow-Ups
- Review upcoming events
- Reminders at the right time
- Prepare recurring schedule summaries
Multi-Step Agentic Work
- Delegate research plus browser interaction
- Organization and follow-up included
- Work too long for a single chat
Scheduled tasks are managed from a dedicated task view where you can review active, scheduled, and completed work in one place.
| Pattern | Trigger | Output | Best For |
|---|---|---|---|
| Recurring check | Schedule | Summary or alert | Daily briefings |
| Monitoring | Schedule or event | Meaningful updates | Topic tracking |
| App workflow | Task assignment | Cross-app actions | Team operations |
| Research | Goal + task | Updated findings | Long projects |
After scheduled work finishes, a Dot can proactively surface results through configured channels. Set one primary channel per Dot to keep updates readable instead of scattered.
The Cloud Computer and Availability
Each Dot gets a dedicated cloud computer with its own browser environment. Cloud work continues independently of your local machine, which is what makes long multi-step workflows practical. You can open the computer view to inspect what the Dot is doing and step in when manual review is useful.
Access depends on plan, workspace configuration, and rollout status:
| Access Path | Status | Notes |
|---|---|---|
| Desktop web | Supported | Primary surface for creating and managing Dots |
| Desktop app | Supported | For users included in the rollout |
| Mobile | Rollout-dependent | Varies by account and stage |
| Business workspace | Admin-managed | Admins control member access |
| Enterprise | Controlled rollout | Governance and connected-tool restrictions apply |
The dedicated cloud computer is separate from optional access to your own machine. Keeping agent execution in the cloud makes permissions easier to reason about and keeps sensitive local access opt-in.
Permissions and Pro Controls
Permissions define the boundary within which a Dot operates. Pro users and admins share responsibility for keeping those boundaries tight.
| Control | Scope | What It Governs |
|---|---|---|
| Connected app permissions | User and workspace | Which apps and services a Dot may access |
| Cloud computer access | Dot capability | Browser-based multi-step tasks |
| Local computer access | Optional user capability | Interaction with your own device |
| Autonomy settings | User configuration | How independently the Dot acts |
| Workspace availability | Administrator | Whether members can use Dots at all |
| Enterprise controls | Enterprise admin | Governance and restrictions |
Pro Readiness Checklist:
- Define one clear ongoing goal per Dot
- Connect only the apps the Dot truly needs
- Set autonomy boundaries and approval rules
- Review the activity dashboard weekly
- Audit connected permissions when projects change
Treat Dot app connections like API credentials. When a project ends, revoke connections that are no longer needed rather than leaving standing access in place.
FAQ
Q: What is an OpenAI Dot?
An OpenAI Dot is a persistent, always-on AI assistant powered by GPT-6 Astra. Each Dot has its own cloud computer and browser, connects to tools like Slack and email, and keeps working toward a goal across many sessions instead of ending when a chat closes.
Q: How is a Dot different from a normal ChatGPT conversation?
A normal conversation ends when you stop typing. A Dot is organized around persistent goals, scheduled tasks, and connected apps, so it can continue work between conversations and proactively follow up through channels like ChatGPT, email, or Slack.
Q: Where can I create and manage OpenAI Dots?
Desktop web and the supported ChatGPT desktop experience are the primary surfaces for creating and configuring Dots. Mobile access varies by account and rollout stage, while business and enterprise workspaces are managed by administrators.
Q: What should I do first for the best OpenAI Dots pro experience?
Start with a single Dot, a clear ongoing goal, tight autonomy boundaries, and only the app connections it needs. Review its activity dashboard regularly, then expand into scheduled tasks and multi-app workflows once the first goal runs smoothly.