How to Set Up OpenAI Dots: Always-On AI Agents

OpenAI Dots launched on September 29, 2026, as always-on personal agents powered by GPT-6 Astra. Unlike a standard ChatGPT chat, a Dot keeps working toward goals you assign, uses its own cloud computer, and connects to thousands of apps. You stay in control through rules, approvals, and activity reviews.

This guide shows you how to create your first Dot, connect tools, set boundaries, and assign a practical first task. Access rolls out gradually to eligible Pro and Business Premium users.

What OpenAI Dots Actually Do

Dots run continuously in the background. They research, draft, update materials, and prepare work for your review. Each Dot has its own cloud computer and browser. You can open that computer anytime to see what it is doing.

You reach your Dot through ChatGPT on desktop, web, and mobile after setup. It also works in Slack and Microsoft Teams. Context carries across channels. Conversations with your Dot do not count toward regular ChatGPT usage limits.

OpenAI includes your first Dot in the Pro or Business Premium plan at no extra cost. Enterprise, Edu, and Healthcare workspaces need an admin to enable the beta. Availability excludes the European Economic Area, Switzerland, and the UK for personal Pro accounts.

Check Eligibility and Prepare

Confirm you are on a supported plan. Pro users outside the EEA, Switzerland, and UK can access Dots as the rollout reaches them. Business Premium users get access across supported regions. Enterprise users ask their admin to turn on the beta.

You must create the Dot on desktop. Mobile creation is not supported. Download the ChatGPT desktop app for macOS or Windows, or use ChatGPT in a desktop browser. Visit chatgpt.com/dots once the feature appears in your account.

Decide on a first responsibility before you start. Good starting points include reviewing your calendar each morning, drafting weekly updates from notes, or preparing materials from connected documents. Clear goals help the Dot know when to stop and ask for input.

Create and Personalize Your Dot

Open ChatGPT on desktop and start the Dot setup flow. Give your Dot a name. The default handle is @yourname-dot. Naming it changes the handle to @yourname-agentname. You can edit the name or avatar later from the profile.

Choose a character or pet for its appearance. The Dot may also generate one for you. After setup, it introduces itself and begins learning your goals and standards from ChatGPT memory and the context you provide.

You can pause the Dot at any time from the profile menu. Resetting deletes the Dot, its conversations, memories, and scheduled tasks. You then create a new one from desktop.

Connect Apps and Optional Computer Access

Connected apps let the Dot read information and act within permissions you set. Open the Dot’s profile, go to Customize, then Plugins. Select a supported app and review what it can do. Permissions are shared with other ChatGPT features.

OpenAI states Dots can connect to over 4,000 apps through the plugin ecosystem. Common choices include calendar, documents, email, and project tools. Disconnecting an app does not erase information the Dot already learned. To clear that, reset the Dot.

Your Dot starts with its own cloud computer. Local computer access is optional and starts off. In the desktop app, open the profile, find Computers, and choose Allow access for your machine. Confirm the prompt. The computer must stay online with the app open for local work.

Local tasks appear separately in the sidebar. You can revoke access later. Certain sensitive actions, such as changing a password, always stay with you.

Set Rules, Approvals, and Oversight

Dots follow built-in rules that decide when to act alone and when to ask. You add Custom Rules under Settings or the profile’s Customize section. For each action, choose one behavior:

• Take action without asking
• Take action if pre-approved (you explicitly requested it)
• Ask before taking action
• Hand off to you

Built-in safety requirements always apply. Auto-review checks actions that could affect accounts or share information against your rules and safeguards. Monitoring can pause or stop work if it detects a concern.

Review progress in Activity View. The profile shows In progress, Scheduled, and Completed work. You can redirect the Dot or take over its cloud computer when needed, then return control.

Proactive research uses read-only tools on connected apps while you are away. It cannot send messages or change content without your rules allowing it. Review consequential outputs yourself. Dots can still make mistakes.

Assign Your First Task and Schedule Work

Describe the goal and the outcome you want. Share the sources or files it needs. For example: “Review my calendar for the next three days. Flag conflicts and prepare a short summary of priorities. Ask me before changing any events.”

Attach files with the + button in the conversation. The Dot works on its cloud computer or connected tools and brings results back for review. Reply in the thread to give corrections or more context.

Ask it to schedule recurring work, such as a morning calendar check or a weekly draft. Manage scheduled tasks in the profile or the Scheduled section. You can change timing, pause runs, or set notifications.

Start with drafts and research rather than actions that send messages or spend money. Expand permissions only after you see consistent results.

Tips for Better Results

Give a clear finish line. Tell the Dot what “done” looks like and when to stop and ask. Name the sources of truth, such as a specific document or connected calendar.

Use Custom Rules early for high-stakes actions. Keep local computer access off until you need it for a specific workflow. Check Activity View regularly, especially for background and scheduled work.

Memory helps the Dot learn your preferences over time. It receives ChatGPT memories and can form its own from connected apps. Reset if you want a clean start.

If access is missing, confirm your plan, region, and that you opened chatgpt.com/dots on desktop. Rollout is gradual, so it may take days to appear.

Watch a Hands-On Setup Demo

This video walks through creating a Dot, exploring the interface, and testing early capabilities.

Conclusion

OpenAI Dots shift ChatGPT from reactive chats to ongoing agents that prepare work while you focus elsewhere. Setup takes a few minutes on desktop. Connect only the apps you need, set clear rules, and begin with reviewable tasks.

The feature is still rolling out, so check your account for access. As you use it, the Dot learns your standards and can take on more structured responsibilities. For related agent workflows, see our guide on using computer-use tools for browser tasks.

Sources

• Introducing dots — OpenAI, September 29, 2026
• Getting started with your dot — OpenAI Help Center
• OpenAI safety and privacy details for Dots, linked from the announcement