- OpenAI Dots are always-on AI agents powered by GPT-6 Astra, announced at DevDay on September 29, 2026
- Each Dot has a dedicated cloud computer and browser for independent long-running work
- Dots connect to 4,000+ apps and follow up via ChatGPT, email, Slack, and SMS
- Availability depends on ChatGPT plan, workspace settings, and rollout status
- Setup takes five steps: create, name, define goal, configure autonomy, assign first task
What Is the OpenAI Dots Personal Assistant?
The OpenAI Dots personal assistant is an always-on AI agent product announced by OpenAI at DevDay on September 29, 2026. Unlike a normal ChatGPT conversation that ends when you close the tab, a Dot is a persistent agent that keeps working toward assigned goals. It is powered by GPT-6 Astra, OpenAI's flagship model designed for complex reasoning, coding, computer use, and research.
Every Dot runs on its own cloud computer with a browser, can connect external tools, and reports back through channels like ChatGPT, email, Slack, and SMS. OpenAI has also added a dedicated Dots section to its official safety documentation, reflecting how significant this product is to the company's agent strategy.
| Feature | Detail |
|---|---|
| Release date | September 29, 2026 (DevDay) |
| Underlying model | GPT-6 Astra |
| Agent type | Always-on persistent AI agent |
| Compute | Dedicated cloud computer + browser per Dot |
| App connections | 4,000+ supported apps |
| Follow-up channels | ChatGPT, email, Slack, SMS |
A regular ChatGPT chat answers one prompt. A Dot holds a persistent goal, keeps its own workspace and computer, and continues working between your conversations — that continuity is the entire point.
How to Set Up Your First Dot
Creating your first Dot is straightforward if you approach it with a clear purpose. The setup flow lives inside ChatGPT on desktop web or the supported desktop app, and the decisions you make during creation shape how independently the Dot operates.
Start a New Dot
Open the Dots experience in ChatGPT on desktop web or the supported desktop app and choose to create a new dot.
Choose a Descriptive Name
Name the Dot after its responsibility — a project, workflow, research topic, or recurring operational task.
Define the Goal
Write the ongoing outcome the Dot should work toward, not a single isolated prompt. A focused goal gives it stable direction across work sessions.
Configure Autonomy and Boundaries
Set instructions for what the Dot should prioritize, how independently it should act, and when it should ask for approval.
Attach Context and Files
Add documents and reference material so the Dot has the project context it needs.
Open the Activity Dashboard
Review completed actions, follow current tasks, and manage ongoing configuration from the Dot's activity view.
Set autonomy boundaries before assigning sensitive work. A Dot that can act between your conversations is powerful, but approval rules decide when it should pause and ask you first.
Cloud Computer and Connected Apps
Each Dot works inside its own dedicated cloud computer. This environment keeps running independently of your local machine, which makes it ideal for longer multi-step workflows like research, browsing, and file organization. You can open the Dot's computer view to inspect its current state and step in manually whenever needed.
Connected apps extend what a Dot can reach. Launch materials repeatedly highlight the 4,000+ app connections figure as a core selling point, with each connection following the permissions granted by the user or workspace.
Cloud Computer
- Dedicated remote environment per Dot
- Runs independently of your device
- Inspectable, with manual intervention possible
Slack
- Team communication context
- Uses authorized workspace access
- Enables channel-aware workflows
- Works with approved messages
- Limited to granted account permissions
- Handles follow-ups and summaries
Files
- Project documents and references
- Only explicitly provided or authorized files
- Gives the Dot persistent context
Workspace Apps
- External information and actions
- Controlled by connected-app settings
- Extends workflows across services
Channels
- Notifications and interaction paths
- Requires supported channel configuration
- ChatGPT, Slack, email, SMS
| Access Type | What It Means |
|---|---|
| Cloud computer | Dot's own remote environment for browser and multi-step tasks |
| Local computer access | Optional, separate capability involving your own device |
| Connected apps | Granted per user or workspace permission |
| Communication channels | Configured paths for updates and task exchange |
The dedicated cloud computer is separate from optional local-computer access. Cloud execution happens remotely and safely isolated; local access is an opt-in capability you control independently.
Scheduled Tasks and Practical Use Cases
Scheduled work is what turns a Dot from a chat partner into a genuine personal assistant. You can assign work for a specific future time, set it to repeat, and receive proactive updates when results are ready — no new prompt required.
| Use Case | How the Dot Helps | Best For |
|---|---|---|
| Ongoing research | Collects, organizes, and updates findings across sessions | Evolving topics |
| Calendar review | Reviews upcoming events, prepares recurring summaries | Busy schedules |
| Reminders and follow-ups | Surfaces important tasks at the right time | Deadlines |
| Recurring monitoring | Repeatedly checks changing information | Market or news tracking |
| Connected-app workflows | Combines actions across permitted services | Cross-tool tasks |
| Multi-step agent work | Research, browsing, organization, follow-up in one flow | Long projects |
Create a Scheduled Task
Tell the Dot what should happen and when — a one-time future task or a recurring schedule.
Run Recurring Checks
Let it review updates, check connected information, or prepare recurring summaries automatically.
Receive Proactive Updates
Results arrive when scheduled work finishes, without starting a new conversation.
Review Task Activity
Open the task management view to see active, scheduled, and completed work.
Choose a Dot when a task needs continuity — recurring monitoring, evolving research, or workflows spanning multiple apps. For a one-off question, a normal ChatGPT conversation is still the faster option.
Availability, Permissions, and Controls
Access to Dots depends on your ChatGPT plan, workspace configuration, rollout stage, and platform. Advanced configuration is centered on desktop experiences, while mobile access varies by account and rollout.
| Access Category | Status | Notes |
|---|---|---|
| ChatGPT access | Plan-dependent | Tied to supported plans and current rollout |
| Desktop web | Supported | Full create, configure, and manage capability |
| Desktop app | Supported | Available for users in the rollout |
| Mobile | Rollout-dependent | Varies by account and platform |
| Workspace access | Admin-managed | Admins control member availability |
| Enterprise | Controlled rollout | Workspace-level governance and restrictions |
Permissions define the boundaries within which every Dot operates:
| Control | Scope | What It Governs |
|---|---|---|
| Connected app permissions | User and workspace | Which apps the Dot may access |
| Cloud computer access | Dot capability | Browser-based, multi-step remote work |
| 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 |
| Enterprise controls | Enterprise admin | Governance and restrictions |
Before diving into setup, confirm that Dots are available on your ChatGPT plan and region. Availability is rolling out gradually, and workspace administrators may need to enable access for teams.
Before You Deploy Your First Dot:
- Confirm Dots availability on your ChatGPT plan
- Choose a focused goal with clear boundaries
- Set autonomy and approval rules
- Connect only the apps the Dot truly needs
- Review the activity dashboard after the first task
FAQ
Q: What is the OpenAI Dots personal assistant?
OpenAI Dots are always-on AI agents announced at DevDay on September 29, 2026. Powered by GPT-6 Astra, each Dot has its own cloud computer and browser, connects to thousands of apps, and keeps working toward assigned goals between your conversations.
Q: How is a Dot different from a normal ChatGPT conversation?
A normal conversation ends when you stop prompting. A Dot holds a persistent goal, maintains its own workspace and cloud computer, runs scheduled tasks, and proactively follows up through ChatGPT, email, Slack, or SMS.
Q: Do I need to keep my computer on for a Dot to work?
No. Each Dot runs on a dedicated cloud computer that works independently of your local machine. You can inspect the computer view and step in manually when needed, but execution continues in the cloud.
Q: Can I control what my Dot is allowed to access?
Yes. Permissions cover connected apps, cloud and optional local computer access, autonomy settings, and workspace-level availability. Enterprise administrators can apply additional governance and restrictions.
Q: Where can I create and manage Dots?
Dots can be created and managed through ChatGPT on desktop web and the supported desktop app. Mobile access varies by account, platform, and rollout stage, and workspace admins control availability for teams.