OpenAI Dots: Always-On AI Agents That Work While You Sleep — Features, Pricing & Honest Guide (2026)

Open laptop glowing at night while an OpenAI Dot AI agent works on tasks automatically

An always-on AI agent keeps working on your tasks even when your laptop lid is closed and you're asleep. (Photo via Bitrix24)

Imagine assigning your AI assistant a task at 11 PM, going to sleep, and waking up to find it finished: the research compiled, the proposal drafted, the spreadsheet updated, and a summary waiting in your Slack. That is the promise of OpenAI Dots, announced on September 29, 2026, at OpenAI's DevDay. Dots are always-on AI agents that work continuously in the background toward goals you define — even when you are not using ChatGPT.

This is one of the biggest shifts in how we use AI since chatbots went mainstream. In this guide, I break down exactly what Dots are, how they work, what they cost, how they compare to Meta's Muse, and — most importantly — whether they are actually worth paying for in 2026.

What Are OpenAI Dots?

An OpenAI Dot is a personal AI agent that runs on GPT-6 Astra, OpenAI's most capable model. Unlike a regular ChatGPT conversation, which starts fresh every time you open the app, a Dot is persistent: it has memory of your goals, keeps working between conversations, and acts independently within boundaries you set.

Each Dot gets its own cloud computer and browser. That means it can open websites, fill out forms, read documents, and work across apps the way a human assistant would — not just answer questions in a chat window. OpenAI describes Dots as a move from "AI you prompt" to "AI you delegate to."

Sam Altman, speaking at DevDay, said the experience changed his own workflow: he described getting some of his attention back and no longer feeling "addicted to his phone" the way he used to. That quote captures the pitch: Dots are meant to remove the constant back-and-forth of prompting and checking.

How Do Dots Work?

Here is the flow, step by step:

  1. You create a Dot in the ChatGPT desktop app or desktop browser, give it a name, and describe the ongoing goal — for example, "track competitor pricing and send me a weekly summary" or "monitor our support inbox and draft replies."
  2. You connect your apps. Dots can connect to more than 4,000 applications through integrations, so the agent can pull data from your tools, read documents, and update systems.
  3. You set Custom Rules. This is the control layer: you define what the Dot may do on its own, when it must ask for approval first, and what it should never do. Dots also get an automatic review of their planned actions.
  4. The Dot works in the background. It runs on its cloud computer, keeps going when you log off, and reaches you through Slack or Microsoft Teams (text messaging support is coming soon). When it needs a decision on a sensitive action, it comes back to you.

In one official demo, a Dot spotted a software bug alert in Slack and started investigating it on its own — digging through code, identifying the issue, and preparing a fix for review. That example shows the intended use: ongoing, multi-step work that would normally sit in a human's queue.

Key Features of OpenAI Dots

1. Persistent, Always-On Operation

Dots do not sleep. They keep working toward your defined goals 24/7 and carry context across sessions, so you never have to re-explain the project. This is the core difference from a chatbot: it is continuity, not conversation.

2. Own Cloud Computer and Browser

Each Dot runs on a dedicated cloud computer with its own browser. It can navigate websites, download files, and use web apps — real computer use, not just API calls. OpenAI says these run on isolated cloud computers with privacy safeguards for business customers.

3. 4,000+ App Integrations

Dots connect to thousands of apps through integrations and plugins, plus Slack and Teams messaging built in. They can work across your stack: research in the browser, update a doc, message your team, log a result.

4. Custom Rules and Approval Gates

You stay in charge. Custom Rules let you set three levels of behavior: act without asking, act only with prior approval, or never act. Dots are also trained not to handle certain sensitive tasks themselves — OpenAI says things like changing passwords or moving money between accounts always stay in the user's hands.

5. Proactive Research

Beyond reacting to your requests, a Dot can do read-only proactive research: watch a topic, gather information over time, and surface what matters. For a freelancer or small business, think "monitor competitor launches" or "track industry news on this niche."

6. Team and Specialist Dots

OpenAI is building specialist Dots for specific functions, and is working with Microsoft to bring them into Agent 365 — its business agent platform. Enterprise, Edu, and Healthcare workspaces can try a beta version once their admin enables it.

OpenAI Dots Pricing: What Do They Cost?

Here is the good news: your first Dot is included in eligible plans at no extra cost. There is no separate "Dots price." According to launch reporting, the eligible plans are:

PlanMonthly PriceDots Access
ChatGPT Pro 100$100First Dot included
ChatGPT Pro 200$200First Dot included
ChatGPT Pro 500$500First Dot included
Business Premium$125 per seatFirst Dot included
Free, Go, PlusVariesNot eligible at launch

A few important pricing details:

  • Dot usage does not count toward your ChatGPT usage limits during the first month. Tasks a Dot starts in Codex or ChatGPT Work do count.
  • OpenAI plans to let users add more Dots later and scale each Dot's speed or workload — but has not published pricing for extra Dots yet.
  • Availability is region-limited: Pro subscribers in the European Economic Area, Switzerland, and the UK are excluded from the initial rollout. Business Premium users can access Dots in all supported ChatGPT regions. OpenAI attributes the gap to strict regional regulation.

OpenAI Dots vs Meta Muse: The Race for Always-On Agents

Dots launched just weeks after Meta's Muse (launched September 8, 2026), Meta's own personal AI agent. The two products are clearly racing toward the same idea — persistent agents that act on your behalf — but they take very different routes to get there:

FeatureOpenAI DotsMeta Muse
LaunchedSeptember 29, 2026September 8, 2026
Underlying modelGPT-6 AstraMuse Spark
Where it runsOwn cloud computer + browserMuse Secure VM (dedicated cloud computer)
PricingIncluded in Pro ($100+) and Business Premium plansFree tier, plus $20 and $100/month paid plans
InterfacesChatGPT web, desktop, mobile; Slack; Teams; texting sooniOS, Android, muse.ai, WhatsApp
Target userProfessionals, teams, developersConsumers first, expanding to work use
Safety controlsCustom Rules, auto-review, safety monitorSeparate Sentinel agent approves outbound actions

The split is clear: Muse bets on mass consumer adoption with a free tier and WhatsApp, while Dots starts with professionals who pay for performance. Muse is positioned for everyday personal tasks — shopping, booking travel, managing calendars. Dots is positioned for real work: projects, research, documents, and workflows that span apps.

Who Should Actually Use Dots?

Dots are not for everyone — and OpenAI's pricing makes that obvious. Here is my honest breakdown:

A great fit if you are:

  • A freelancer or agency owner juggling client research, reporting, and admin. A Dot could compile weekly reports, monitor competitors, or triage inboxes while you do the creative work.
  • A small team lead who lives in Slack and Teams. Dots meet you where you work and can keep projects moving between meetings.
  • A developer or technical founder. The bug-triage demo shows where Dots shine: watching systems, investigating issues, and preparing fixes for human review.
  • A marketer running repetitive research loops — trend monitoring, campaign reporting, or content research pipelines.

Probably overkill if you are:

  • A casual ChatGPT user. At $100/month minimum, Dots only make sense if the agent genuinely saves you billable hours every week.
  • Someone in the EEA, Switzerland, or the UK on a Pro plan — you are excluded from the rollout for now.
  • Looking for a free or cheap personal assistant. Muse's free tier is the more sensible starting point for everyday personal tasks.

The Honest Take: Safety, Limits, and Whether It's Worth It

No review of always-on agents would be complete without the safety question. Giving an AI the power to act across your apps is genuinely powerful — and genuinely risky. OpenAI is open about this: in its safety materials, the company admits an agent's mistakes can reach beyond a chat window, such as editing the wrong file or sharing private information. The Custom Rules, auto-review, and permissions system are designed to contain those failure modes — but as of launch, no independent security audit of the Dots architecture had been published.

My practical advice, regardless of which agent you use:

  • Start with read-only tasks. Research, monitoring, and drafting are high-value and low-risk. Let the Dot watch and report before you let it touch anything.
  • Connect only the apps a task needs. A Dot that monitors competitors does not need access to your bank account.
  • Run a short pilot. Give one Dot one responsibility for a week, review what it did, tighten the rules, and only then expand.
  • Keep approval gates on anything irreversible. Publishing, deleting, spending money, and messaging people should always pass through you.

Is it worth $100/month? If a Dot reliably saves you even two or three hours of repetitive work per week, the math works for most professionals. But do not pay for the hype: the value is in specific, ongoing workflows, not in "having an AI agent" as a novelty. And if you are a casual user, Muse's free tier covers the personal-assistant use cases without the bill.

How to Set Up Your First Dot Safely

  1. Open the ChatGPT desktop app or desktop browser (Pro or Business Premium required, in an eligible region).
  2. Create your first Dot, give it a name, and write one clear responsibility — not five.
  3. Connect only the apps that responsibility actually needs.
  4. Set Custom Rules: what it may do alone, what needs your approval, and what it must never touch.
  5. Run it for 48 hours, review its activity log, refine the instructions, and then decide whether to scale.

Frequently Asked Questions

What is an OpenAI Dot?

An OpenAI Dot is an always-on AI agent inside ChatGPT, powered by GPT-6 Astra, that works continuously in the background on tasks and goals you define. Each Dot has its own cloud computer, connects to thousands of apps, and reaches you through Slack or Teams.

How much do OpenAI Dots cost?

Your first Dot is included in ChatGPT Pro ($100–$500/month) and Business Premium ($125/seat/month) plans at no extra cost. Pricing for additional Dots has not been announced yet.

Are OpenAI Dots available in my country?

As of October 2026, Dots are rolling out in eligible markets but exclude Pro subscribers in the European Economic Area, Switzerland, and the UK. Check your ChatGPT account and OpenAI's help pages for current availability in your region.

What is the difference between OpenAI Dots and Meta Muse?

Both are always-on AI agents, but Dots start with paid ChatGPT plans and target professional workflows across apps, while Muse offers a free tier and focuses more on everyday consumer tasks through Meta's ecosystem.

Can a Dot act without asking me?

Yes — within the Custom Rules you set. You define what it may do alone, what needs prior approval, and what it should never do. Sensitive actions like changing passwords or moving money always stay with you.

Do Dots learn my preferences?

Yes. Dots carry context across sessions and learn your preferences over time, which is why they improve the longer they run on a consistent responsibility.

Conclusion

OpenAI Dots mark a genuine turning point: AI is moving from something you talk to into something you delegate to. The always-on design, the cloud computer, the 4,000+ integrations, and the approval-gate controls add up to the most serious attempt yet at a true AI coworker.

But the honest bottom line is this: Dots are a professional tool with a professional price. If you have repetitive, ongoing workflows that eat your week, a well-configured Dot could pay for itself many times over. If you are a casual user, the free alternatives will serve you better — and the agent race is only just getting started, so waiting costs you nothing.

What is the first task you would trust a Dot with? Let me know in the comments.

Post a Comment

Previous Post Next Post