ChatGPT Dots — also searched as OpenAI Dots or simply Dots — are OpenAI’s always-on AI agents inside ChatGPT. You give a Dot a goal; it keeps working toward that goal on its own cloud computer, connects to thousands of apps, and asks before the risky moves. OpenAI announced them at DevDay on 29 September 2026 in Introducing dots.

Think of Dots as Codex energy for everyday work: cute blobby avatars on the outside, workplace software that can use other software on the inside — closer to a coworker than a personal shopper.

At KnowAnt, we treat agent launches like any other brief: name the product, say what it actually does, list who can get it, and keep the hype in a separate drawer. Pair this page with GPT-6 Astra explained (the model under Dots) and Meta Muse AI Agent (the consumer rival). For prompt-driven chat that is not agentic, use the AI prompts list and Write with AI.

Quick Answer: What Are ChatGPT Dots?

Dots are OpenAI’s always-on agents in ChatGPT, powered by GPT-6 Astra. Each Dot has its own cloud computer and browser, can connect to 4,000+ apps via plugins, and keeps making progress between chats.

Table of Contents

• Quick Answer: What Are ChatGPT Dots?
• What Are ChatGPT Dots?
• Dots vs a Normal ChatGPT Chat
• Key Features
• How ChatGPT Dots Work
• Pricing, Plans & Who Gets Access
• ChatGPT Dots in India vs Meta Muse
• Dots vs Meta Muse
• Realistic Limits in the Wild
• How to Get Started
• Privacy, Safety & Custom Rules
• Risks and Realistic Limits
• Frequently Asked Questions
• Who Should Try Dots First
• Conclusion
• Sources

What Are ChatGPT Dots?

A Dot is a named, personalized agent that lives inside ChatGPT. OpenAI’s pitch: frontier intelligence that has your back — learns what matters to you, works on your behalf, and takes important work off your plate so you get time and attention back.

Unlike a normal ChatGPT thread that ends when the reply ends, a Dot can:

• take ongoing responsibility for a task
• keep making progress between conversations
• run on a dedicated cloud computer (with optional access to your desktop via the ChatGPT desktop app)
• use plugins / apps you connect (email, calendar, files, creative tools like Blender and GIMP in early demos)
• show up in Slack (and Teams in limited alpha), with voice calls and texting on the roadmap

Today you start with one primary Dot, give it a name and look, and make it yours. OpenAI’s longer vision is teams of Dots working together — plus specialist Dots for enterprise roles (access management, IT-provisioned hardware, deep integrations with systems of record).

Dots vs a Normal ChatGPT Chat

In short: Dots are task-oriented, always-on, and tool-connected, with human control on send / pay / delete-class actions.

Key Features

Always-on cloud computer. Your Dot clicks around a virtual machine while you watch in a second pane — same broad pattern as Muse’s Secure VM, but OpenAI’s packaging leans work, not WhatsApp errands.

GPT-6 Astra as the brain. Dots ride OpenAI’s computer-use flagship. For model limits, coding, and API pricing, see GPT-6 Astra explained.

4,000+ apps via plugins. Connectors inherit the permissions you already gave those plugins — drafting email does not automatically mean sending email.

Desktop computer access. Through the ChatGPT desktop app, a Dot can use your machine (one at a time, off by default) for local files, video edits, deploys, and other “on my laptop” work. Early reviews say this is where Dots feel strongest.

Channels. ChatGPT app, voice call, Slack; Microsoft Teams in invite-only alpha; iMessage / RCS waitlists for some Pro users.

Custom Rules. Four behaviors: take action without asking, take action when you say so, ask before taking action, or hand off to you. Password changes are always handed to you.

Specialist / enterprise preview. Identity for access management, IT hardware, Agent 365–style controls with Microsoft, and role-tuned specialist Dots (accounting, legal, marketing) on the DevDay stage.

How ChatGPT Dots Work

The loop in plain English:

1. Goal — you assign ongoing work (“monitor #bugs and draft fixes,” “rebuild this site,” “keep the launch checklist moving”).
2. Plan — the Dot breaks the goal into steps on Astra.
3. Act — it uses the cloud browser, plugins, Codex / ChatGPT Work hooks, or your connected PC.
4. Ask when needed — Custom Rules + automatic review gate sensitive steps.
5. Continue — progress keeps going after you close the chat or hang up a call.
6. Report — you get updates in ChatGPT (and Slack if connected).

That is agentic AI applied to ChatGPT’s distribution — not a research demo that only books a restaurant in a sandbox.

Pricing, Plans & Who Gets Access

OpenAI’s launch framing (29 September 2026):

There is no separate Dot SKU today. You pay for the ChatGPT plan; the first Dot is bundled. Plans include an allowance for deeper work (with extended limits in the first month after launch). Later, OpenAI says you will add more Dots and scale speed or monthly work volume — prices for that not announced yet.

Practical takeaway from reviews: early consumer hands-ons used the ~$100/month Pro tier. That gated rollout is intentional: cute avatar, enterprise-grade access bar.

Confirm live plan names and regions in OpenAI’s help / introducing dots page — rollouts move weekly.

ChatGPT Dots in India vs Meta Muse

The practical question for many readers is which agent you can actually use from India.

So for many Indian professionals and students on ChatGPT Pro, Dots is the agent you can try; Muse is the agent you can only read about. A VPN does not unlock Muse if the account region is checked server-side. Always verify in your own ChatGPT / Muse app — availability changes.

Students who need writing and study help without a Pro agent: Learn with AI and Write with AI.

Dots vs Meta Muse

Both products launched in the same agent wave (cute avatars, cloud VMs, background work). The fit differs.

If you want the full Muse architecture (Spark, Secure VM, Sentinel), read Meta Muse AI Agent explained.

Realistic Limits in the Wild

Early use of Dots is mixed — strong on controlled work, weaker on open-web consumer errands.

Where Dots often struggle

• Human “press and hold” CAPTCHAs and looping login checks
• Consumer checkout without a polished virtual-card payment path
• Missing a free option and jumping to a paid plan by mistake
• Sites that block OpenAI’s cloud browser footprint

Where Dots often shine

• Redesigning a personal website from a long voice brief
• Editing desktop video clips for social after local computer access
• Deploying site changes while you step away and get phone updates
• Completing errands once a human helps with logins or verification

The pattern: Dots are strongest when the environment is under your control (your site, your files, your permissions). They are weaker as a frictionless personal shopper — that lane fits Muse better today.

How to Get Started

1. Confirm you are on Pro, Business Premium, or an Enterprise workspace with Dots enabled.
2. Open ChatGPT on a desktop app or desktop browser (not mobile web for first-time setup).
3. Start the Dots introduction: name it, style the avatar, get your @handle.
4. Connect a few read-mostly apps first (calendar, email, files). Hold payments and delete rights.
5. Set Custom Rules on day one: ask before send/post; hand off payments and password changes.
6. Optionally connect one personal computer in the desktop app for local work.
7. Add Slack (or wait for Teams alpha) if your team lives there.

If the Dot is missing: wait for rollout lag, check region/age, check plan, or ask an Enterprise admin to flip “Use dots.”

Privacy, Safety & Custom Rules

OpenAI’s control story for Dots:

• Plugins keep their existing scopes — the Dot does not silently widen them.
• Automatic review checks instructions, plugin permissions, Custom Rules, and built-in safety before account-affecting actions.
• Proactive background research is described as read-only (cannot send, edit, or drive browser/computer by itself).
• Password changes are always handed to you.
• Local computer access is opt-in and permission-heavy — grant least privilege.

None of that makes an agent risk-free. Prompt injection in webpages/email and over-broad Drive write access remain the classic failure modes. Start narrow.

Risks and Realistic Limits

Security checks. Cloud-browser agents trip bot walls more often than humans. Budget time for takeovers.

Wrong conclusions. Agents can miss the free option and pick the paid one. Read confirmations.

Money. Early Dots reviews did not show Muse-like one-tap protected checkout. Do not paste bank numbers into a VM you do not trust.

Cost. Bundled with Pro is still a high monthly bar for casual use. Many people will keep Plus + prompts until Dots reach their plan.

Over-trust. Treat the Dot like a junior coworker with system access: powerful when scoped, expensive when trusted blindly.

Region policy. Pro exclusions (EEA / UK / Switzerland) and Muse’s US/Canada-first map will keep changing — re-check before you buy a plan “for the agent.”

Frequently Asked Questions

What is ChatGPT Dots / OpenAI Dots?

Dots are OpenAI’s always-on AI agents inside ChatGPT. Powered by GPT-6 Astra, each Dot has a cloud computer, connects to apps, and can keep working toward goals between conversations.

When did OpenAI launch Dots?

29 September 2026, at OpenAI DevDay. The company published Introducing dots the same day.

Is ChatGPT Dots free?

No separate free Dot on Free/Plus at launch. Your first Dot is included with ChatGPT Pro or Business Premium at no extra Dot fee. Enterprise is admin-gated beta.

How much does ChatGPT Dots cost?

You pay for the ChatGPT plan (Pro reviews often cite the $100/month Pro tier). Extra Dots and higher work quotas are future paid options OpenAI has previewed without final prices.

Can I use ChatGPT Dots in India?

Eligible Pro and Business Premium users in supported markets — including India on OpenAI’s Pro-eligible country list — can get Dots as rollout reaches their account. Confirm in ChatGPT, and check OpenAI’s Introducing dots page for the latest eligible-market notes.

Is Meta Muse available in India?

Not as of early October 2026. Muse launched in the US (then Canada). From India, Dots on an eligible ChatGPT plan is the practical agent option; Muse is not.

How is Dots different from Meta Muse?

Dots lean work / computer use / paid ChatGPT; Muse leans personal life / WhatsApp / freemium. See the comparison table above and Meta Muse explained.

What model powers Dots?

GPT-6 Astra — OpenAI’s computer-use flagship. Details: GPT-6 Astra explained.

Can Dots use my own computer?

Yes, optionally, through the ChatGPT desktop app — one machine at a time, off by default, with many permissions. That is where early reviewers saw the best results (video edits, site deploys).

Can Dots order food or shop for me?

Sometimes, with friction. Hands-ons report CAPTCHAs, blocked cloud browsers, and manual payment entry. Dots can do consumer errands, but that is not the sweet spot compared with Muse-style shopping agents.

How many Dots can I have?

One primary Dot at launch. OpenAI’s roadmap includes multiple Dots and specialist Dots for teams.

Are Dots the same as Codex?

No. Codex remains the coding-focused agent stack. Dots reuse that “do the work while you look away” idea for a wider set of goals — think of it as Codex for broader everyday work, not only code.

Who Should Try Dots First

Good early fit: ChatGPT Pro or Business Premium users who already live in Slack, run side projects or small businesses, and want background help on sites, docs, research loops, and local files.

India Pro users: if Muse is blocked for you, Dots are the agent experiment you can actually run — start with read-only apps and Custom Rules.

Wait-and-see: Free/Plus users, anyone who mainly wants cheap personal shopping help, and people who will not babysit CAPTCHAs or review approvals.

Writers and students: use Dots (if you have Pro) for research chores; still draft the human-facing copy in Write with AI and keep study workflows in Learn with AI.

Conclusion

ChatGPT Dots are OpenAI’s answer to the personal-agent wave: always-on, Astra-powered, cute on the surface, and serious about work underneath. They ship first to high-tier ChatGPT plans, shine when you give them controlled environments (your computer, your site, your plugins), and stumble more on CAPTCHA-heavy consumer websites — exactly what early reviews found.

If you are comparing agents from India, the short version is blunt: Dots can be reachable on eligible ChatGPT plans; Muse is not in India yet. Treat either agent like a junior with system access — connect less than you can, approve what you read, and disconnect what you do not need.

For the model under the hood, read GPT-6 Astra explained. For Meta’s rival, read Meta Muse AI Agent. For everyday writing while you evaluate agents, use Write with AI.

Sources

• OpenAI — Introducing dots