OpenAI Dots are always-on AI agents launched on September 29, 2026 at OpenAI’s DevDay. Each Dot runs on its own cloud computer and browser, powered by GPT-6 Astra, and works toward your goals in the background without step-by-step prompting. You name your Dot, connect it to 4,000+ apps, and message it from ChatGPT, Slack, or Microsoft Teams.
What Is an OpenAI Dot?
Most AI tools wait for you to type something. You prompt, they answer, and the conversation ends. A Dot flips that around: you give it a goal — not a prompt — and it keeps working toward it continuously, in the background.
The biggest shift is the dedicated hardware. Every Dot gets its own isolated cloud computer and browser that you can open at any time to see exactly what it is doing. Close the chat window and the work goes on.
OpenAI calls Dots “remarkably capable, always-on agents built to handle everything.” If the agent concept is new, our guide on what an AI agent actually is covers the basics first.
How Do OpenAI Dots Work?
The model. Every Dot runs on GPT-6 Astra, OpenAI’s flagship model. It plans the work, breaks your goal into steps, and executes them — browsing, using apps, calling tools — on its own.
The cloud computer. Each Dot gets an isolated cloud machine with its own browser, and you can open an activity view to watch it work. By default, background work on your apps is read-only: nothing is sent, edited, or deleted without permission.
The connections. Through OpenAI’s plugin ecosystem a Dot can reach more than 4,000 apps. You message it in ChatGPT on desktop, web, and mobile, or in Slack and Microsoft Teams, with context carried across channels. Text-message support is coming soon.
You also set Custom Rules: let it act freely, make it ask first, or block actions entirely. Password changes and money transfers always stay in your hands.
What Can a Dot Actually Do?
OpenAI’s own examples are practical: a developer deploys a dedicated Dot to monitor customer feedback and implement bug fixes as new reports come in; a scientist has a Dot rerun data analysis as fresh experimental data arrives.
Real early users report the same diligence: in one outside test, an agent noticed its owner had forgotten to bill a publication, prepared the invoice, and waited for approval before sending it. OpenAI has used Dots internally for months — when a bug is reported in the company’s Slack, agents start investigating on their own. OpenAI eventually envisions teams of Dots working together for one person.
Who Can Use Dots and How Much Do They Cost?
Dots are rolling out to ChatGPT Pro and Business Premium subscribers in eligible markets. Your first Dot is included at no extra cost, and usage does not count against plan allowances for the first month.
Details worth knowing: Enterprise, Edu, and Healthcare customers can try a beta if a workspace admin enables it; Pro subscribers in the EEA, Switzerland, and the UK are excluded at launch; and pricing for extra Dots has not been announced.
The same week, Meta expanded its Muse agent to small businesses (Shopify, Dropbox, Slack, Canva, Notion, Stripe, and Zoom integrations, plus a free tier), and Nvidia released an Open Agent Safety Platform with 100+ partners. Persistent agents have clearly moved from concept to product.
How Do You Set Up Your First Dot?
If you have an eligible plan, setup is straightforward:
- Name your Dot. Give your primary Dot a name and a cartoon avatar — the bubbly branding is meant to make always-on agents feel approachable.
- Connect only the apps it needs. Link the tools it should work with through the plugin connections. Start with as few as possible.
- Write your Custom Rules. Decide what the Dot may do freely, what it must ask about first, and what is blocked. Start strict: require approval for anything external — emails, messages, payments.
- Give it one clear goal. Pick a single, bounded, recurring task — “summarize every new customer support ticket each morning and draft replies” is ideal. Measure the hours it saves before expanding.
Which Tasks Are Best for a Dot?
Dots earn their keep on recurring, well-defined background work: monitoring customer feedback and support tickets, hours-long research, drafting from transcripts and voice notes, and flagging unusual patterns in data or spend.
Keep away from autonomous agents for now: payments, irreversible deletions, sensitive data access, or messages sent in your name — at least until you have watched a Dot behave correctly for weeks. Give each agent its own credentials and only the app access it needs.
How Do You Talk to a Dot Effectively?
A Dot is not a search box — it performs best with a written brief. State the goal, the constraints, and what “done” looks like. “Track mentions of our product on Reddit this week, summarize the top 10 complaints in a table, and flag anything urgent” will outperform “check Reddit.”
The prompting fundamentals still apply — our guide on how to write better AI prompts explains the method that works for agents too. Review the activity view regularly in the first weeks: Dots learn from your feedback, and correcting a wrong assumption early teaches it your preferences.
For a hands-on comparison with a free agent that runs on your own machine, see our guide to DeepSeek Harness — the open-source alternative to cloud agents.
Should You Get a Dot Right Now?
If you are already paying for ChatGPT Pro or Business Premium, trying your first Dot is an easy yes — it is included, the first month of usage is free, and the guardrails are genuinely thoughtful.
If you are not on one of those plans, there is no reason to upgrade purely for Dots yet. The feature is days old, extra dots are not priced, and the safety review of always-on agents is still playing out. Meta’s Muse offers a free tier, and DeepSeek Harness costs nothing at all.
AI is moving from something you chat with to something that works while you sleep.
Frequently Asked Questions
What is the difference between OpenAI Dots and ChatGPT?
ChatGPT is a chat interface: you send prompts and get answers. Dots are always-on agents with their own cloud computers that pursue goals in the background and connect to over 4,000 apps.
How much does an OpenAI Dot cost?
Your first Dot is included with ChatGPT Pro and Business Premium, and usage does not count against plan allowances for the first month. Pricing for extra Dots has not been announced.
Who can use OpenAI Dots?
ChatGPT Pro and Business Premium subscribers in eligible markets. Pro subscribers in the EEA, Switzerland, and the UK are excluded at launch; Enterprise, Edu, and Healthcare customers can try a beta if their admin enables it.
Can a Dot access my computer or send messages for me?
Only with your permission. Background work on connected apps is read-only by default, a Dot cannot touch your own devices unless you allow it, and Custom Rules can make it ask first or block actions.
Sources
- TechCrunch — “OpenAI launches Dots, its bubbly agentic avatar” (September 29, 2026): https://techcrunch.com/2026/09/29/openai-launches-dots-its-bubbly-agentic-avatar/
- Jeffrey Stop — “Always-on agents reach SMBs as safety and governance tools arrive” (October 5, 2026): https://jeffreystop.com/news/2026-10-05-1140-ai-tech-news/
- TechCEO Daily — “OpenAI Dots: always-on ChatGPT agents with cloud computers”: https://www.techceodaily.com/ai/openai-dots-always-on-agents
Leave a Reply