ChatGPT Dots: What It Is and How to Use It

In short. ChatGPT dots are always-on agents that live in your ChatGPT account. A dot has its own cloud computer, connects to your apps through ChatGPT plugins, and keeps working on a goal between conversations. You create one in ChatGPT, give it a name and a job, connect the apps that job needs, set rules for what it may do without asking, and then reach it in ChatGPT, Slack or Teams. This guide follows OpenAI's dots announcement and the dots feature page, both checked on November 11, 2026.

The two most-watched tutorials show the same setup in action and are worth a look:

• ChatGPT Dots Tutorial for Beginners by Kevin Stratvert, about 85,000 views. He builds a YouTube coach dot, connects vidIQ and checks its results the next day. The video is sponsored by OpenAI, as he says at the start.
• How to Master ChatGPT Dots in 12 Minutes by Tenex, about 47,000 views. It sends a dot into existing ChatGPT projects and has it supervise a Codex build.

What are ChatGPT dots?

OpenAI introduced dots on September 29, 2026 as "always-on agents built to handle everything". Three things set a dot apart from a chat:

• Its own computer. Each dot works on its own cloud computer and browser, so it keeps going when your laptop is closed. You can open that computer to inspect its work, and you can also let it use your own laptop.
• It learns you. A dot starts with context from your ChatGPT memory and learns your preferences from feedback over time.
• It owns a goal. You hand it a responsibility, and it works out the next steps, keeps making progress between conversations and brings you work to review.

Kevin Stratvert's tutorial puts the difference simply: a chat answers when asked, a scheduled task repeats one prompt on a timetable, and a dot takes the bigger goal and figures out what to work on next.

How do you set up your first dot?

1. Create it. OpenAI says to create your first dot in the ChatGPT desktop app or a desktop browser. In Kevin Stratvert's walkthrough, dots sits in the upper left of ChatGPT, just below New chat. If you don't see it, dots may not be on your account yet: OpenAI is rolling them out across eligible markets, and in Enterprise workspaces an admin turns them on.

2. Choose where it works. By default the dot uses its own cloud computer. You can also give it permission to use your laptop, which lets it work with files on your machine.

3. Name it. OpenAI's launch post says you can give your primary dot a name and make it your own. Kevin names his "YouTube coach" and picks a look for it. After setup you can message it from the ChatGPT mobile app too.

How do you give a dot a job to own?

4. Give it a goal, not a list of steps. Tell it the outcome and how you want updates. Kevin's prompt asks for 20% more daily views over 90 days and progress reports every Monday and Thursday, then leaves the plan to the dot. It came back with a baseline, where his traffic was coming from and a list of video ideas.

5. Point it at the context it already has. If you have lots of ChatGPT projects, Tenex's tip is to send the dot into them to learn what you are working on. When it can't find a thread on its own, copy the thread's deep link and paste it in.

6. Review and steer. On the feature page OpenAI says you can see what your dot is working on, give feedback, change direction or pause it at any time. Background work shows up in Activity View.

How do you connect your apps to a dot?

7. Connect only what the job needs. Dots reach apps through ChatGPT plugins, and OpenAI says they can connect to over 4,000 apps. You add and manage them in ChatGPT's Plugins tab. Kevin searches for vidIQ there, authorizes it, then types @vidIQ in his prompt so the dot uses it.

Plugin permissions are shared across dots, ChatGPT, ChatGPT Work and Codex, so anything you already connected is available to your dot within the permissions you granted.

8. Set your rules. Dots come with built-in rules for when to act and when to ask. Custom Rules let you allow specific actions, require approval or block them, for example telling your dot never to send emails. Actions that could affect your accounts or share information go through auto-review, and steps such as changing a password always stay with you.

When you aren't working with it, a dot does what OpenAI calls proactive research: it reads your connected apps with read-only tools that can't send messages or change content.

Can you message a dot in Slack or Teams?

Yes. OpenAI says your dot is within reach through ChatGPT, Slack or Teams, and you can also call it when you want to talk something through. The suggested flow is to start a project in ChatGPT, share context with the working team in Slack, and let your dot follow through. It carries context across every channel. Texting your dot is listed as coming next.

You can also set up a separate Slack account for your dot, so it has its own identity there rather than posting as you.

How do teams share work with dots?

Today a dot works on behalf of the person who owns it. For organizations, OpenAI is previewing specialist dots: each gets its own identity, credentials and access to the systems it needs for a set role, such as invoice processing or customer support. These start as focused enterprise pilots, and OpenAI is working with Microsoft to manage them in Agent 365. OpenAI also says that over time it envisions teams of dots working together.

Mistakes the tutorials make easy

• Memory goes both ways. Your dot receives ChatGPT memories and can add to them. Turning off Memory in ChatGPT stops the sharing but doesn't delete what the dot already received.
• You can't edit single memories yet. OpenAI's FAQ says you currently can't view, delete or change individual dot memories. Deleting the dot starts deletion of its context.
• Disconnecting an app isn't deleting. Removing a plugin stops new access but doesn't remove what the dot already learned from it.
• Review consequential work. OpenAI's own line: dots can still make mistakes, so check anything that matters before it goes out.

Where does Kylon fit?

A dot is one agent working for one person, reachable from the tools that person uses. Some teams want one place where several people and several agents work the same request in the open. That is how Kylon works.

• Rooms and threads. People and agents join a Room for one area of the business, such as an account or a launch. Each request runs in a thread every member of the Room can read, and anyone can pick it up.
• Several agents, each with a role. A research agent, a writing agent and a data agent can work the same thread, each with its own skills and connections. Claude Code and Codex can join as members too.
• Approval. When an agent prepares an email, a record change or a post that leaves the workspace, it stops at a draft the person who asked reviews.
• Automations. Recurring work runs on a schedule, a webhook or a record change, and its results land in the Room.

For a side-by-side view, see OpenAI dots alternatives. For how shared agents work across tools, see multiplayer AI agents for teams, compared.

Facts are from OpenAI's Introducing dots post, the ChatGPT dots feature page and the dots privacy, security and safety FAQ, checked on November 11, 2026. View counts are from YouTube on the same day. ChatGPT, OpenAI and dots are trademarks of their respective owners and are used here for identification only.