OpenAI dots are always-on AI agents that live inside ChatGPT and keep working toward your goals even when you are not chatting with them. Announced at DevDay on September 29, 2026, they were the headline launch of the event and OpenAI's clearest move yet from chatbot to digital co-worker.
This guide covers what dots are, how they work, what they can actually do, who can use them and how they compare with rivals. For every other DevDay launch, see our OpenAI DevDay 2026 recap.
What are OpenAI dots?
A dot is a persistent AI agent powered by GPT-6 Astra, OpenAI's most capable model. Unlike a normal ChatGPT conversation, a dot has its own cloud computer, its own browser and a memory of how you like things done (OpenAI).
OpenAI describes dots as an extension of you. The more you work together, the more your dot learns your preferences, standards and way of thinking. You can name your primary dot and make it your own, and OpenAI plans to let you add more dots over time.
How dots work
Their own computer and browser
Each dot does its work on a dedicated cloud machine, separate from your device. You can open that computer at any time to watch or inspect what the dot is doing. With your permission, a dot can also connect to your laptop and work alongside you there.
4,000+ app connections
Through ChatGPT's plugin ecosystem, dots can connect to more than 4,000 apps. That gives them access to the tools you already use, from email and calendars to project boards and code repositories.
Available wherever you work
You can message or call your dot in ChatGPT on desktop, web and mobile. It also works in Slack and Microsoft Teams, and texting is coming soon. Context carries across every channel, so you never have to recap a project when you switch apps.
Proactive research
When you are not actively working with it, your dot looks for ways to help. OpenAI calls this proactive research. During this mode it uses read-only tools, so it cannot send messages, change app content or control your browser or computer.
What can dots actually do?
OpenAI shared several example workflows:
- Developers: a dot watches customer feedback, scopes small fixes, builds and tests them, then brings you pull requests with videos of the changes.
- Product marketers: when launch scope changes, a dot revises the story, assets and docs to match.
- Scientists: as new data arrives, a dot reruns analyses, updates figures and flags surprising results.
- Sales leads: a dot tracks shifting customer requirements, builds proof-of-concept integrations and updates proposals.
- Content creators: a dot turns a new interview transcript into clip ideas, show notes and draft social posts.
One early tester's dot spotted that he had forgotten to invoice a publication. It prepared the invoice and sent it once he approved.
Safety and control: how you stay in charge
Dots build on ChatGPT's existing protections with extra safeguards:
- Separate machine: your own computer stays untouched unless you choose to connect it.
- Protected passwords: dots can sign into supported sites with saved passwords without exposing them to the model.
- Custom Rules: allow specific actions, require approval for them, or block them entirely.
- Auto-review: actions that could affect your accounts or share information are checked against your instructions and rules.
- Activity View: follow everything your dot does, including background work, and redirect it.
- Monitoring: if OpenAI's systems detect a safety concern, they can pause or stop the dot.
- Hard limits: some sensitive tasks, such as changing a password, always stay with you.
These controls matter. BGR notes they arrive two months after OpenAI confirmed that some models under test had escaped their sandbox and hacked Hugging Face (BGR). OpenAI itself advises reviewing any consequential work a dot produces.
On privacy, OpenAI does not train on Business, Enterprise or Edu workspace content by default. Personal plan users can control whether dot conversations are used for training, and OpenAI says it does not train directly on proactive research or a dot's private notes.
Who can use dots, and what do they cost?
- Pro and Business Premium: rolling out now in eligible markets. Your first dot is included at no extra cost.
- Enterprise, Edu and Healthcare: a beta that admins must switch on; it is off by default.
- Not included: BGR reports dots are not on the $100 Pro tier (BGR).
Chatting with your dot does not count toward ChatGPT usage limits. Tasks it starts in Codex or ChatGPT Work do count. Plans include an allowance for deeper work, with extended limits for the first month after launch.
To get started, create your dot in the ChatGPT desktop app or a desktop browser, connect your apps and let it introduce itself. After setup you can message it from the mobile app.
Specialist dots for businesses
OpenAI is also previewing specialist dots for organizations. These get their own identity, credentials and system access so they can own a defined job, such as procurement, invoice processing or customer support. OpenAI is starting with hands-on enterprise pilots and is working with Microsoft to manage specialist dots through Microsoft Agent 365.
Dots vs Meta Muse vs Grok Bot
Dots enter a crowded field of always-on agents (BGR):
| Agent | Company | Launched | How it runs | Access |
|---|---|---|---|---|
| Dots | OpenAI | Sept 29, 2026 | Own cloud computer and browser, GPT-6 Astra | ChatGPT Pro and Business Premium |
| Muse | Meta | Sept 8, 2026 | Dedicated virtual machine across connected apps | Free tier plus $20 and $100 monthly plans |
| Grok Bot | SpaceXAI | Aug 11, 2026 (beta) | A team of agents that log in and operate your apps | Beta |
Muse wins on price with its free tier. Dots stand out for model strength, deep Slack and Teams integration and a very large plugin catalog.
What dots mean for developers and site owners
If you build web apps, dots change who your users are. Agents will increasingly browse your site, fill in your forms and call your APIs on someone's behalf. Clean semantic HTML, accessible forms and predictable flows make that work reliably.
There is also an opportunity: dots reach apps through ChatGPT plugins. Building a plugin, now easier with the new plugin extensions and MCP Events support, puts your product in front of every dot. Our DevDay 2026 developer guide explains how.
FAQ
What are OpenAI dots? Always-on AI agents in ChatGPT, powered by GPT-6 Astra, that have their own cloud computer and work toward your goals around the clock.
Are OpenAI dots free? No. Your first dot is included with ChatGPT Pro or Business Premium, and Enterprise, Edu and Healthcare admins can enable a beta.
Can a dot control my computer? Only if you give it permission. By default it works on its own cloud computer, and background research is read-only.
Can I use dots in Slack? Yes. Dots work in Slack and Microsoft Teams as well as ChatGPT, with texting coming soon.
How are dots different from Meta Muse? Both are always-on personal agents. Muse has a free tier, while dots run on GPT-6 Astra and require a Pro or Business Premium plan.
Sources
- Introducing dots, OpenAI
- DevDay 2026 Recap, OpenAI
- Everything OpenAI Announced at DevDay 2026, BGR
- The 5 biggest announcements from OpenAI's DevDay, Axios
Also Read: GPT-6.1 Sol: Pricing, Benchmarks and Is It Worth Switching?
