OpenAI launched dots on September 29, 2026. If you already use Grok Bot, you probably want to know two things: what a dot is, and whether it does the same job. This note sticks to what OpenAI and SpaceXAI have published. Botski has not used a dot yet, so nothing here is a hands-on test.
Botski is independent. It is not xAI, SpaceXAI, OpenAI, or Cursor.
What a dot is
OpenAI's launch post calls dots “remarkably capable, always-on agents built to handle everything.” In plain terms:
- Model: dots are powered by GPT-6 Astra.
- Computer: each dot works on its own cloud computer with its own browser. You can open that computer to check its work, and you can let your dot connect to your laptop if you choose.
- Apps: through OpenAI's plugins, dots can connect to more than 4,000 apps.
- Where you talk to it: ChatGPT on desktop, web, and mobile, plus Slack and Microsoft Teams. You can also call your dot. OpenAI says texting is coming soon.
- Background work: when you are not working with it, a dot does what OpenAI calls “proactive research” using your connected apps with read-only tools. OpenAI says those tools can't send messages, change app content, or control your browser or computer.
- Controls: you pick which apps it can reach. Custom Rules let you allow, require approval for, or block specific actions. Auto-review checks actions that could affect your accounts, and some steps, such as changing a password, always stay with you.
Who can get one
According to OpenAI, dots are rolling out in ChatGPT to Pro and Business Premium users in eligible markets. Enterprise users, including Edu and Healthcare, can try the beta when their workspace admin turns it on. Your first dot is included in a Pro or Business Premium plan at no extra cost. Talking to your dot doesn't count toward ChatGPT usage limits, but tasks it starts in Codex or ChatGPT Work do.
For now you start with one primary dot. OpenAI says teams of dots are a future plan, and it is previewing specialist dots for enterprises, including work with Microsoft on Agent 365 (OpenAI, InfoWorld). You create your first dot in the ChatGPT desktop app or a desktop browser, then you can message it from mobile.
Side by side with Grok Bot
| ChatGPT dots | Grok Bot | |
|---|---|---|
| How many | Start with one primary dot; more dots later (OpenAI) | Create several Bots and give each a job (dot.com) |
| Computer | Each dot works on its own cloud computer (OpenAI) | All your Bots share one persistent cloud computer. Isolation is per user, not per Bot (FAQ) |
| Where you talk to it | ChatGPT (web, desktop, mobile), Slack, Teams, voice calls; texting soon | Desktop app on macOS, Windows, or Linux, plus iOS and Android (FAQ); Team Bots also in Slack (docs) |
| Keeps working when you leave | Yes, with read-only “proactive research” when idle | Yes. Closing the app or laptop doesn't stop a background turn or routine (FAQ) |
| Approval controls | Custom Rules (allow, require approval, block) plus Auto-review | Auto Review with Ask first and Allow automatically rules; Ask first wins (docs) |
| Passwords | Can use saved passwords without exposing them to the model | Hands you the computer to type passwords and codes yourself (docs) |
| Teams | Specialist dots in enterprise preview; teams of dots planned | Team Bots: one shared Bot, private chat per teammate (docs) |
| Plans | ChatGPT Pro, Business Premium; Enterprise beta | Paid individual Cursor plans, Cursor Teams, or a linked SuperGrok, SuperGrok Plus, or SuperGrok Heavy subscription (Cursor Docs, mobile docs) |
Where they overlap
Both products are built around the same idea: an agent with its own computer that keeps working after you close the app and stops for your approval on risky steps. Coding shows the overlap well. OpenAI's launch post describes a dot that turns customer feedback into tested fixes and brings you complete PRs with attached videos. A day later, the official @bot account said Grok Bots can hand off coding tasks to Cursor, manage PRs with the GitHub and Origin plugins, and share video demos of what they build.
Where they differ
- One agent or a roster. OpenAI starts you with one primary dot that can juggle several projects. Grok Bot is built around a roster of named Bots, each with its own job, that can message each other and share a computer.
- One shared computer or one per dot. Grok Bot's FAQ says not to treat separate Bots as a security boundary, because they share files, browser sessions, and logins. OpenAI says each dot works on its own cloud computer.
- Where you already pay. Dots come with ChatGPT Pro and Business Premium. Grok Bot comes with Cursor and SuperGrok plans. If you already pay for one of those, that may decide it for you.
Meta's Muse belongs in the same group. It is a personal agent that runs on its own secure cloud VM, and you can talk to it in the Muse app or in WhatsApp. TechCrunch noted that the cartoon look of dots is similar to Muse.
What we don't know yet
- How dots perform day to day. We haven't used one.
- What extra dots will cost. OpenAI says you will be able to add more dots “in the future.”
- When texting arrives and when dots reach more plans.
We will update this note when we can test it ourselves or when OpenAI or SpaceXAI publish more.
Independent, not xAI
Botski is an independent site for people learning Grok Bot. It is not official xAI, SpaceXAI, OpenAI, or Meta documentation, and none of those companies endorse it. Official Grok Bot docs: docs.x.ai/grok-bot.