ChatGPT dots guide (October 2026): what OpenAI's always-on agent does, which plans get it, how usage counts and how to set one up

What a ChatGPT dot is, in OpenAI's words

OpenAI calls dots "always-on agents in ChatGPT that can take on ongoing work and keep making progress between conversations". Everything below comes from OpenAI's release notes, Help Center, pricing page and announcement, read on October 21.

• Its own computer. Each dot runs on GPT-6 Astra with a sandboxed cloud computer and browser, so it can research, analyse, create files and run tools while you are away. You can open that computer from its profile, watch, or take over and hand control back.
• Your apps and memory. It works through the apps you connect (OpenAI says over 4,000 through plugins) and starts with context from your ChatGPT memory, then forms its own memories as it works.
• Ongoing responsibility. You give it a goal rather than a single prompt. It works out next steps, runs reminders and recurring checks you schedule, and can do background "proactive research" with read-only tools to spot ways to help.
• You stay in charge. You decide what it may do alone, what needs approval and what it must hand back. Decisions that need your judgment come back to you.
• Where you reach it. Message or call it in ChatGPT on web, mobile and desktop; message it in Slack or Microsoft Teams. Texting is a limited beta for US Pro users only.

How it differs from a normal chat: a conversation ends when you stop typing, while a dot keeps a task open, comes back with progress and remembers context across days. It also differs from a Codex or ChatGPT Work task, though it can start those for you.

Which plans and markets get dots

• Age: 18 or older. OpenAI says dots are not available to users under 18.
• Rollout: gradual; access "may take several days" to reach an account. Check at chatgpt.com/dots with the right account and workspace selected.
• Country: beyond the Pro exclusions above, ChatGPT itself is offered only in OpenAI's supported countries.

What this means if you buy from us: a top-up or a ready-made account gives you the official Pro plan. Whether a dot shows up on that account is decided by OpenAI's age, market and rollout rules; we cannot switch it on.

How dots usage counts against your plan

OpenAI has made three statements so far. Read together:

OpenAI's release note of September 29 says dots usage does not count toward Pro, Business and Enterprise allowances "for the next month"; by our reading that runs to about October 29. After that OpenAI will "share usage terms for each plan", and as of October 21 it has not published them. OpenAI also says you will later be able to add more dots and scale each one's speed or monthly workload, which suggests extra capacity will be sold separately; prices have not been announced.

Why Pro 200 can matter for a dot: the dot is the same on every Pro tier, but the Codex and Work tasks it starts come out of your plan allowance, and Pro 200 has more of it than Pro 100. From October 30 every Pro 200 moves to the new, lower allowance, so plan on that when your dot starts Codex or Work tasks. OpenAI publishes no per-plan figures; your Settings → Usage page shows the real limit.

How to set up your dot, step by step

1. Check that your account qualifies: Dots need a Pro plan (any of the three tiers) or a Business Premium seat, and the user must be 18 or older. Pro access excludes the EEA, Switzerland and the UK at launch, and rollout is gradual, so it can take several days to reach an account. Open chatgpt.com/dots while signed in to see whether yours has it.
2. Create the dot on a supported surface: Use the ChatGPT desktop app (macOS or Windows), chatgpt.com in a desktop browser, or the latest ChatGPT app on iOS or Android, and follow the onboarding prompts. Mobile web is not supported.
3. Name it: Its handle starts as @yourname-dot; give it a name and it becomes @yourname-agentname. You can pick a character or a pet as its avatar and change both later from its profile.
4. Connect the apps it needs: Open its profile → Customize → Plugins and review each app's permissions before connecting. Connections are shared with ChatGPT, ChatGPT Work and Codex. Slack and other messaging channels are set up from desktop.
5. Set Custom rules: Under Customize → Custom rules, choose per action: take action without asking, take action if pre-approved, ask before taking action, or hand off to you. Pre-approved means you asked for that action explicitly in your prompt.
6. Give it a goal and check its work: Describe the outcome and attach files with +. Ask for recurring checks (for example a morning calendar review). Follow its work under In progress, Scheduled and Completed in its profile, or in Activity View in the desktop app; pause it from the ••• menu at any time.

Optional extras: to let it work on your own machine, open its profile in the ChatGPT desktop app on that computer, find Computers → Your computer and select Allow access (it stays off until you do; connecting a computer to Codex does not count). It can also create tasks in Codex cloud environments you have already set up. Reset (••• → Reset) deletes the dot with its conversations, memories and scheduled tasks.

What OpenAI says dots are good for

These are OpenAI's own examples from its Help Center and dots page, not our promises:

• Presentations: pull together data and customer feedback, investigate what changed and keep a deck updated as new figures arrive.
• Content: turn an interview transcript into show notes and social posts, then carry your edits across the drafts.
• Apps: scope a change, write code and test it, so you get working changes to review; or watch customer feedback and build small fixes.
• Sales and deals: check customer requirements against emails and product docs and prepare an evaluation plan.
• Engineering projects: map dependencies for an API migration, prepare code and tests, and track remaining calls.
• Everyday follow-through: review your calendar each morning, flag a clash, research a topic, remind you later.

Writers and analysts will find the research and drafting side covered in more depth in ChatGPT Pro 200 for research and writing; developers in ChatGPT Pro 200 for coding.

Controls, limits and privacy

• Custom rules. For supported actions, choose: take action without asking, take action if pre-approved, ask before taking action, or hand off to you. They cannot switch off OpenAI's built-in safeguards.
• Steps that stay with you. Changing a password or transferring money between financial accounts is always handed back. Permanently deleting data, installing or running software from an unrecognised source and granting new security-sensitive access need confirmation each time. Purchases with a card saved on a merchant's site need your approval.
• Auto-review. Before actions such as sending an email or changing files, a separate system checks the step against your instructions, rules and safety requirements; for an email it checks the recipient and message.
• Memory. The dot receives your ChatGPT memories and makes its own. You cannot view or edit single dot memories; you delete them only by resetting the dot. Disconnecting an app stops new access but does not delete what the dot already learned. Its context does not keep credentials, images or screenshots.
• Training. On personal plans, your "Improve the model for everyone" setting decides whether dot conversations and work may be used to train models; Business, Enterprise and Edu are not used by default. OpenAI notes human review can still happen in limited cases, including safety, even with that setting off.
• Launch limits. One dot per plan for now; no mobile web; it cannot call you; it cannot have its own standalone email address; it can make mistakes, and prompt-injection protections reduce the risk without removing it. Turn on multi-factor authentication, as OpenAI advises.

Which plan to buy for dots

Top-up or ready-made? A dot "starts with context from ChatGPT memory" and uses your connected apps, so the account you already use is the better home for it: the top-up adds Pro to that account, and your chats, memory and plugin connections stay. A ready-made account starts empty, which suits people with no account or who want a separate work identity for the dot. Either way it is for one person: a dot connects your email, calendar and Slack, and OpenAI's terms forbid sharing credentials.

Who should not buy Pro for dots

• You live in the EEA, Switzerland or the UK. Pro does not include dots there at launch. Business Premium does, or wait for OpenAI to expand.
• You are under 18. Dots are not available to you.
• You need a fixed price for heavy dot use. The terms after the first month are not out yet, and extra dots or capacity look set to be sold separately.
• You only want reminders and scheduled prompts. Plus already has scheduled tasks, at ¥185 ≈ 27.58 USDT.
• You will not connect any apps or review its work. A dot is only as useful as the access you give it, and OpenAI says to check important results.
• You want to text it from outside the US, or have it call you. Neither is available at launch.

Getting Pro for your dot here

1. Sign in to the account you want the dot on and check Settings → My Plan. Free, Plus and Pro accounts all qualify for Pro 100, 200 or 500 top-ups (new subscription, upgrade or renewal).
2. Pick the tier on the pricing page and check the live price and stock.
3. Paste the session JSON the order page asks for. It is checked before you pay; if the check fails you pay nothing. Your password stays with you.
4. Pay in USDT (TRC20 or BEP20), exactly the amount shown. Each order covers 30 days.
5. Wait for activation, usually 5 to 30 minutes, on the same account. Then open chatgpt.com/dots.

No account? A ready-made Pro 200 (¥1,349 ≈ 201.05 USDT, ≈ 7.31 USDT more than the top-up in the first month): pay, enter a delivery email, click the Cloudflare verification email, and our team delivers the account by hand in 30 minutes to about 4 hours. Change the initial password as soon as it arrives. To renew, message Telegram @ccnb1937 with your order ID and pay the top-up price. Full walkthrough: Ready-made ChatGPT Pro 200 account.

Can you trust us? What is in writing

• Checked before you pay. A top-up checks your session first, and you pay nothing if it fails; the data is used once and cleared after delivery.
• Refund terms. If an OpenAI regional policy change makes the service unavailable, you get a refund for the unused days, processed within 1 business day and returned in USDT. Bans caused by breaking OpenAI's usage policy are not refunded. Contact the bottom-right live chat promptly if anything goes wrong.
• Who we are. ClaudeMax has operated since November 2024 and has served more than 20,000 users (our own figure). We are not OpenAI and not an authorized OpenAI reseller; OpenAI decides dots availability and usage terms.
• One person per account. OpenAI's terms prohibit sharing credentials and reselling access, and OpenAI may restrict access. More in Is ClaudeMax legit?