ChatGPT Dots are persistent AI agents created by OpenAI that run inside an isolated virtual cloud computer, allowing them to browse the web, use connected apps, and complete multi-step tasks in the background without requiring you to keep a chat window open.
If you use AI primarily as a reactive tool where you type a prompt, wait for a response, and manually move that text elsewhere, OpenAI is trying to alter that routine. Announced at DevDay 2026, these new agents are designed to give chatbots a place to actually execute workflows rather than just answering questions in a browser tab.
By the end of this guide, you will know what ChatGPT Dots can do, how their cloud environment works, what safety boundaries are in place, and whether they fit into your daily routine.
Key takeaways
Here is what you need to know about OpenAI’s latest agent rollout:
- ChatGPT Dots run on isolated cloud virtual machines powered by GPT-6 Astra, enabling them to work continuously in the background (Tom’s Guide).
- They connect to over 4,000 apps via plugins and can be reached through Slack, Microsoft Teams, voice calls, and upcoming SMS support (Tom’s Guide).
- The first Dot is included at no additional charge for ChatGPT Pro subscribers paying $100 a month (Tom’s Guide, Wired).
- Proactive background research operates in read-only mode so agents cannot independently send messages or spend money without approval (Tom’s Guide).
What happened
OpenAI has introduced ChatGPT Dots, described as always-on AI agents that operate independently to handle tedious everyday chores and complex professional assignments (Wired). Powered by the startup’s GPT-6 Astra model, these agents get their own cloud-based virtual machine to browse websites, open applications, and process information while you are away from your desk (Tom’s Guide, 9to5Google). Rather than serving as a single-turn chatbot, a Dot learns your preferences over time and maintains context across multiple communication modes (Wired).
Users can initially set up a Dot using ChatGPT on the desktop web or desktop app, where they assign its purpose and link necessary services (Tom’s Guide). Once established, you can communicate with your Dot via mobile apps, voice calls, Slack, and Microsoft Teams, with text messaging support planned for the future (Tom’s Guide, Wired). CEO Sam Altman also showcased specialist versions tuned for professional fields like accounting, email marketing, and legal analysis during the announcement (Wired).
In the future, if you like, you can work with a whole team of dots.
Sam Altman, CEO, OpenAI
At a glance
This reference table summarizes the core technical and commercial details of the ChatGPT Dots release.
| Detail | What we know |
|---|---|
| Who | OpenAI |
| What | ChatGPT Dots persistent AI agents |
| Underlying Model | GPT-6 Astra (9to5Google) |
| Where | Desktop web, desktop app, mobile, Slack, Microsoft Teams (Tom’s Guide) |
| Price | Included with ChatGPT Pro ($100/mo) and Business Premium (Tom’s Guide, Wired) |
| Availability | Rolling out today for eligible subscribers and enterprise betas (Tom’s Guide) |
| Integrations | Over 4,000 apps via plugins (Tom’s Guide) |
| Security Limits | Read-only mode for background tasks; explicit approval required for sensitive actions (Tom’s Guide) |
For most users, that means powerful background capabilities are currently restricted to high-tier subscribers using desktop configuration.
Understanding the technical architecture behind these agents clarifies why OpenAI structured their security and operational limits the way it did.
Why it matters
ChatGPT Dots represent a fundamental shift away from single-turn chat interfaces toward autonomous background workers. Until now, using generative AI meant opening a tab, entering a prompt, copying the resulting text, and manually executing the next step in another application (Tom’s Guide). Dots attempt to eliminate that friction by letting an AI agent operate continuously inside an isolated cloud virtual machine (Tom’s Guide).
You type a prompt into a chat window, wait for a reply, and manually move data into other applications.
You assign a project to a persistent agent that executes tasks on a cloud virtual machine while you are away.
Administrative chores and research can run in the background, freeing up hours of manual clicking.
This release arrives amid intense competition in the personal agent space, following Meta’s release of its Muse agent (Wired, The Guardian). By positioning Dots within professional channels like Slack and Microsoft Teams alongside specialized versions for legal and accounting tasks, OpenAI is targeting both everyday power users and enterprise clients (Tom’s Guide, Wired).
- Exact long-term pricing for adding additional Dots beyond the single included instance has not been detailed (9to5Google).
- The exact timeline for general public rollout to free or lower-cost subscription tiers remains unannounced (Wired).
- Full SMS text messaging support is planned for a future update rather than day-one availability (Tom’s Guide).
These unknowns highlight that while the core technology is launching today, its broader ecosystem integration will take months to mature.
Hype vs reality
Marketing announcements often paint an idealized picture of autonomous software, but the actual implementation of ChatGPT Dots comes with strict technical boundaries.
The AI can completely run your digital life while you sleep.
Background research runs in read-only mode, and sensitive actions require your explicit approval (Tom’s Guide).
The Dot takes over your personal laptop or desktop computer.
Each Dot operates inside an isolated virtual machine in the cloud, separate from your local files (Tom’s Guide).
Recognizing these operational guardrails helps set realistic expectations for what your agent can safely accomplish.
Examining how these boundaries affect different workflows reveals who will benefit most immediately from the launch.
What it means for you
The impact of ChatGPT Dots varies depending on your daily routine, technical requirements, and subscription tier.
If you subscribe to ChatGPT Pro
You receive one Dot included at no additional charge with extended usage limits during your first month (Tom’s Guide, Wired).
Do this: Open the desktop app to configure your first agent and connect your primary productivity apps (Tom’s Guide).
If you manage business operations
You can access enterprise betas and specialist agents tuned for tasks like accounting and email marketing (Wired).
Do this: Coordinate with your workspace administrator to enable enterprise beta access (Tom’s Guide).
Evaluating your specific tier ensures you do not waste time attempting to access features locked behind paid subscriptions.
Who should care (and who can skip it)
Determining whether this release demands your attention depends entirely on how much administrative friction you handle daily.
Worth your time if you…
- Pay for ChatGPT Pro or Business Premium tiers (Tom’s Guide).
- Spend hours juggling multi-step web research and app administration (Tom’s Guide).
- Collaborate via Slack or Microsoft Teams and want integrated AI agents (Tom’s Guide).
You can skip it if you…
- Only use the free tier of ChatGPT (Wired).
- Prefer simple single-turn question-and-answer interactions (Tom’s Guide).
- Have strict data security policies prohibiting cloud agent plugin connections (Wired).
If you fall into the second group, you can safely ignore this launch without missing out on essential everyday chat features.
For eligible subscribers ready to test the new functionality, setting up an initial agent requires following specific desktop steps.
How to try it
Eligible subscribers can configure their first cloud agent directly through desktop applications starting today (Tom’s Guide).
- Open ChatGPT on Desktop Launch ChatGPT on your desktop web browser or desktop application, ensuring you are logged into a Pro or Business Premium account (Tom’s Guide).
- Create Your First Dot Select the option to initialize a new Dot, and define its primary purpose or project scope (Tom’s Guide).
- Connect Necessary Apps Link the external services and plugins you want the agent to access through its isolated cloud environment (Tom’s Guide).
- Switch to Mobile or Messaging Once initial desktop configuration is complete, communicate with your agent via mobile devices, Slack, or Microsoft Teams (Tom’s Guide).
Initial setup must happen on a desktop interface before mobile or team messaging functions become available (Tom’s Guide).
Monitor my connected project management tool for [SPECIFY PROJECT NAME], summarize any updates regarding [SPECIFY TOPIC], and prepare a draft status report for my approval without sending it externally.
Define clear boundaries in your initial prompt so the agent knows exactly which data sources to check and when to pause for your review.
Applying structured constraints prevents the agent from wandering into irrelevant data silos during background execution.
Do not assume a Dot can automatically execute financial transactions or send external messages while you are away; background tasks operate in strict read-only mode until you approve them (Tom’s Guide).
Respecting these security limitations ensures your automated workflows run smoothly without unintended external actions.
Keeping an eye on upcoming software updates and platform changes will help you adapt as OpenAI expands agent capabilities.
What to watch next
As OpenAI expands its agent ecosystem, several upcoming milestones will dictate how useful these tools become for everyday workflows.
- SMS Support Rollout Watch for the official release of text message support, which will allow you to text your AI assistant directly from mobile devices (Tom’s Guide).
- Multi-Agent Controls Keep track of OpenAI’s expected rollout allowing individual users to manage multiple concurrent agents rather than just one (Wired).
- Tier Expansion Timelines Look for announcements regarding when Dots might trickle down to lower subscription tiers or free users (Wired).
Tracking these developments prevents you from guessing release schedules for unreleased features.
Our take
ChatGPT Dots represent an exciting evolution in how we interact with artificial intelligence, shifting the burden of execution from human to machine. However, because they are currently restricted to high-priced subscription tiers and require careful oversight of security permissions, they are not yet an essential utility for casual users. If you already pay for Pro, setting up your first agent is worth exploring to reclaim hours of administrative research.
You can configure your first desktop agent today if you hold a qualifying subscription (Tom’s Guide).
If you subscribe to ChatGPT Pro, log into your desktop app today and set up a single read-only Dot to automate one repetitive administrative task.
Your next step
- GuideAI models guide
Understand the underlying GPT-6 Astra intelligence powering the new agents.
- Prompt hubBusiness & Operations prompts
Explore workflows and prompts to delegate administrative tasks to your Dot.
- Prompt hubEveryday & Personal prompts
Find practical ideas for delegating personal tasks like meal planning and research.
Questions people ask
What are ChatGPT Dots?
ChatGPT Dots are persistent AI agents created by OpenAI that run inside an isolated virtual cloud computer. Powered by GPT-6 Astra, they browse the web and complete multi-step tasks in the background while you are away from your chat window.
How much do ChatGPT Dots cost?
ChatGPT Dots are included at no additional charge for subscribers to the ChatGPT Pro tier, which costs $100 a month, as well as Business Premium and Enterprise beta users where enabled by administrators.
Can ChatGPT Dots access my personal computer files?
ChatGPT Dots operate entirely inside an isolated cloud virtual machine by default and do not automatically access files sitting on your local laptop unless you explicitly grant permission.
How do I communicate with a ChatGPT Dot?
You can communicate with your ChatGPT Dot through the desktop web or app, mobile apps, voice calls, Slack, and Microsoft Teams, with SMS text messaging support planned for future updates.
Are ChatGPT Dots safe for background tasks?
ChatGPT Dots feature built-in safety guardrails, operating in read-only mode during background research so they cannot independently send messages, delete files, or spend money without your explicit approval.
Sources
- Tom's Guide – OpenAI announces ChatGPT Dots — here's what the new AI agents can doUsed for: Details on cloud virtual computers, app integrations, and background read-only safety limits.Report
- Wired – OpenAI’s Dots Are Always-On AI Agents—and Its Answer to Meta’s MuseUsed for: Information on GPT-6 Astra power, Pro tier pricing, and specialist enterprise agent roles.Report
- 9to5Google – OpenAI launches Dots, new ‘always-on agents’ you can assign tasks toUsed for: Confirmation of DevDay 2026 announcements and plugin ecosystem details.Report
- The Guardian – OpenAI announces ‘dots’ agent after scrapping launch of new AI model over safety concernsUsed for: Sam Altman quotes and context regarding halted GPT-6.1 Astra safety testing.Report
Sources checked . We update this story when details change.
Written with AI assistance from the sources above and reviewed by PromptifyLab before publishing. Spotted an error? Tell us and we will fix it.