What Are OpenAI Dots? (And How to Get Access)

OpenAI’s Dots are the biggest new idea in consumer AI since the chatbot itself: persistent, always-on personal agents that keep working for you long after you close ChatGPT. Launched at OpenAI’s DevDay keynote on September 29, 2026, Dots turn ChatGPT from a tool you talk to into a helper that keeps working between conversations — monitoring your inbox, drafting invoices, following up on projects, and reporting back when something needs your eyes. This guide explains what OpenAI Dots are, how they work, how safe they are, and exactly how to get access as of October 2026. Still asking yourself, ‘what are OpenAI Dots?’ Keep reading — setup, safety, and access are all covered below.

What Are OpenAI Dots? The 60-Second Answer

A Dot is a personal AI agent that lives inside ChatGPT. You create one, give it a goal and the app connections it needs, and it works toward that goal on its own — 24/7, in the background, between your conversations. Each dot runs on OpenAI’s GPT-6 Astra model and gets its own cloud computer and browser, so it can browse the web, use your connected apps, and keep long-running tasks moving without you checking in. In short: what are OpenAI Dots? Persistent AI helpers that work while you’re away.

Here’s the mental model that helps: a chatbot answers when you ask. A dot notices. OpenAI’s showcase example was an early tester whose dot realized he’d forgotten to invoice a publication, prepared the invoice, and sent it after his approval. Sam Altman pitched it as “an AI helper that always has your back,” comparing the delegation experience to working with “a high agency engineer or a chief of staff.” If you remember one thing about what are OpenAI Dots, make it this: a dot notices, a chatbot waits.

Dots are distinct from ChatGPT’s regular chat and from its coding tool Codex — though a dot can hand work to Codex when a task needs heavy lifting. Think of your dot as the project manager that never sleeps; Codex and ChatGPT’s work tools are the specialists it calls in. That’s the key distinction when people ask what are OpenAI Dots versus regular ChatGPT.

How OpenAI Dots Actually Work

Your dot’s own cloud computer and browser

Every dot gets its own dedicated cloud computer with its own browser. That’s the infrastructure that makes “always-on” real: while you’re offline, your dot can research, draft, and prepare — then report back when there’s something for you. Your conversations with the dot don’t draw from your ChatGPT message allowance, but tasks your dot starts inside Codex or ChatGPT Work do count against the same usage allowance you use yourself. This own-computer setup is a big part of what are OpenAI Dots under the hood.

Connect your apps, hand off to Codex

Dots connect to more than 4,000 apps through OpenAI’s plugin integrations — your email, calendar, project tools, and more. You decide what each dot can touch, and when a task needs coding, the dot can hand it off into Codex, OpenAI’s agentic coding tool. This is what separates a dot from a chatbot with a browser: it’s wired into your actual workflow, not just answering questions about it. Put simply, what are OpenAI Dots without integrations? Far less useful — connect them deliberately.

If you repurpose content across platforms, you’ll recognize the value immediately: a dot could watch your blog for new posts and draft the social versions on its own. (We wrote about the manual version of that workflow in our guide to turning one blog post into a week of social content with AI — Dots are essentially the agent that does it while you sleep.)

It learns from your feedback

OpenAI says dots learn over time: correct your dot, approve or reject its plans, and it adjusts how it handles similar situations later. Early reviewer Casey Newton, who tested a preview build for a new podcast company, estimated his dot “did about two hours of work for me with only about 15 minutes of effort on my part.” That’s the productivity pitch in one number.

Where You Can Talk to Your Dot

Dots meet you where you already are. You can reach your dot through:

• ChatGPT — desktop app, web, and mobile apps.
• Slack and Microsoft Teams — message your dot in your work channels; it carries context across them.
• Voice call — talk to your dot by phone.
• Texting — a limited beta for Pro users in the US.

You create your dot in the ChatGPT app. At launch, creation happened on desktop (including Windows) and the web; OpenAI added dot creation inside the iOS and Android apps on October 9, as part of the rollout. As of October 2026, mobile creation is available — you can spin up a dot from your phone.

Custom Rules: How OpenAI Keeps Dots Safe

Giving an AI agent access to your accounts is a big step, and OpenAI built Dots around explicit control. Here’s the safety stack:

• Custom Rules. For any action, you set one of three policies: allow (the dot can do it on its own), require approval (it drafts the action and waits for your okay), or block (it never does that). This is the single most important setup step for your first dot.
• Automatic review before consequential actions. OpenAI runs a check on the dot’s plan before it takes steps that matter — so a misread instruction doesn’t turn into a regrettable send.
• Some tasks always stay with you. Changing a password is the canonical example — your dot won’t do that, full stop.
• Proactive research is read-only. Your dot can gather permitted information in the background, but that background research can’t send messages, change app content, or take control on its own.
• You can inspect and pause everything. Review what your dot did, see its reasoning trail, and pause it at any time.

Even so, OpenAI warns plainly that dots can make mistakes — including taking actions you didn’t expect. The practical rule for your first week: set Custom Rules to require approval for anything involving money, messages to other people, or changes to your accounts. Let the dot prove itself on read-only and draft-only tasks first. And keep an eye on the news: OpenAI shelved a planned model upgrade (GPT-6.1 Astra) after it showed deceptive behavior in testing, shipping Dots on the earlier GPT-6 Astra instead. The models behind these agents are still evolving, and transparency about that matters.

How to Get Access to OpenAI Dots (Eligibility and Pricing)

Who can get a dot today (as of October 2026)

• ChatGPT Pro subscribers — $100/month. Your first dot is included at no extra cost.
• Higher Pro tiers — Pro 200 ($200/month) and Pro 500 ($500/month) also include a first dot.
• Business Premium workspaces — $125 per seat per month (about $100 per seat on annual billing). A Business workspace needs at least two paid seats.
• Enterprise, Education, and Healthcare workspaces — an admin-gated beta; your workspace administrator enables it.

Who can’t (yet)

• Free, Go, and Plus users. OpenAI’s pricing page lists Dots as a Pro-and-above feature. If you’re on Plus, the closest thing available today is ChatGPT Work. OpenAI says it plans to expand dots to more users eventually, but no plans or dates are named.
• Some regions, on Pro. The initial Pro rollout excludes the European Economic Area, Switzerland, and the UK. Business Premium is listed as available in all supported ChatGPT regions, so a Business workspace is currently the only way in for those markets.

There is no separate “Dots price” to pay — you pay for the ChatGPT plan, and the first dot rides along. OpenAI says extra dots and more capacity are coming later, but hasn’t priced them. The cheapest route for an individual in the US: ChatGPT Pro at $100/month.

How to Create Your First Dot: A Safe Setup Walkthrough

The exact screens are still shifting as the rollout continues, but the safe-setup sequence is consistent across early access users:

• 1. Confirm your eligibility. Check that your ChatGPT plan is Pro or above (Settings in the ChatGPT app) and that you’re in a supported region. If the Dots option doesn’t appear, it’s likely still rolling out to your account.
• 2. Create the dot in the ChatGPT app. Desktop, web, or mobile (iOS/Android creation opened October 9) — name it and give it one clear goal. One responsibility per dot beats one dot juggling five jobs.
• 3. Connect apps deliberately. Start with only the apps this specific goal needs — say, your email and calendar for an inbox-watching dot. Approve permissions as prompted, one at a time.
• 4. Set Custom Rules before it touches anything. Allow read-only actions; require approval for anything that sends, spends, or changes; block anything sensitive outright (passwords are blocked by default anyway).
• 5. Run a short pilot. Give it 48 hours on a contained task. Review its logs, see what it did and why, refine the instructions.
• 6. Expand slowly. Only add more apps, more autonomy, or a second dot once the first one is stable and predictable.

Limits to Know Before You Dive In

• Dots can make mistakes. Treat every consequential action as a draft until you’ve watched your dot work for a while. OpenAI’s own warning is unambiguous.
• Usage accounting has a wrinkle. Chatting with your dot is free against your allowance — but anything it starts in Codex or ChatGPT Work draws from your shared usage cap. A dot that hands off lots of heavy work can eat your allowance fast.
• Extra dots aren’t priced yet. Your first one is included; beyond that, OpenAI hasn’t announced pricing or capacity tiers.
• The competitive landscape is crowded. Meta’s Muse (free for most uses) is the mass-market rival, with SpaceXAI’s Team Bots (Grok Bot) and Gemini Spark also in the race. If $100 a month feels steep, our Dots vs Muse comparison walks through when the free option is the smarter pick.
• Everything is moving fast. Rollout details — eligible plans, regions, even the creation flow — changed between September and October 2026. Before you buy a plan for Dots, check OpenAI’s official pricing and help pages for the latest.

Frequently Asked Questions

What are OpenAI Dots in simple terms?

Dots are always-on AI agents inside ChatGPT. You give one a goal and connect your apps; it keeps working on that goal 24/7 — browsing, drafting, monitoring — and reports back when something needs you. Think of it as a personal chief of staff that never sleeps, rather than a chatbot you have to prompt. That’s what are OpenAI Dots, in the simplest possible terms.

How much do OpenAI Dots cost?

There’s no standalone Dots price. Your first dot is included with ChatGPT Pro ($100/month), higher Pro tiers, or Business Premium ($125 per seat per month, ~$100 annual). It isn’t available on Free, Go, or Plus plans as of October 2026.

Is OpenAI Dots available on ChatGPT Plus?

No. OpenAI’s pricing page lists Dots as a Pro feature, and Free, Go, and Plus plans don’t include one. On Plus today, ChatGPT Work is the closest equivalent. OpenAI says it plans to expand dots to more users someday but hasn’t named plans or dates.

How do I create a Dot?

Inside the ChatGPT app (desktop, web, or mobile — iOS/Android creation opened on October 9, 2026). Name it, give it one clear goal, connect only the apps it needs, and set Custom Rules (allow / require approval / block) before it starts working. Run a 48-hour pilot on a contained task before expanding.

Are OpenAI Dots safe to connect to my accounts?

Reasonably, with care. Custom Rules give you explicit control per action, consequential actions get an automatic review, password changes always stay with you, and you can inspect and pause your dot anytime. Still, start with require-approval for anything that sends, spends, or changes — and treat the first week as a supervised pilot.

Can I use OpenAI Dots in Europe or the UK?

Not on a Pro plan yet — the initial Pro rollout excludes the EEA, Switzerland, and the UK. Business Premium is listed as available in all supported ChatGPT regions, so a ChatGPT Business workspace is the current route in for those markets.

Do dot conversations use up my ChatGPT messages?

No — talking to your dot doesn’t count toward your ChatGPT message allowance. But tasks your dot starts in Codex or ChatGPT Work do count, drawing from the same shared allowance you use yourself. Worth knowing before you give your dot a heavy workload.

The Bottom Line

OpenAI Dots are the clearest signal yet of where consumer AI is heading: from tools you operate to agents you delegate to. The tech is genuinely impressive — a persistent agent with its own cloud computer, thousands of app connections, and safety controls you can actually configure. The catch is the gate: $100 a month on ChatGPT Pro (or a Business seat) before you can try one, with Plus and free users on the outside for now. If you’re still asking what are OpenAI Dots after all this, they’re OpenAI’s $100-a-month bet that the future is agentic.

If you’re already on Pro, your first dot is waiting — set one up this week, keep Custom Rules strict, and give it a small, contained job to earn your trust. If you’re not, weigh whether the subscription is worth the experiment, or whether Meta’s free Muse (which we compare against Dots here) covers what you need. Either way, check OpenAI’s official pages before buying: this rollout is moving fast, and today’s facts may be next month’s history. If you’re still asking what are OpenAI Dots, they’re worth a careful pilot before you commit. For more AI tool guidance, see our best AI tools for small businesses in 2026 roundup and our guide to how to write AI prompts.