OpenAI has a new kind of digital assistant, and it comes in the shape of a blob. The company introduced Dots at its DevDay 2026 event in San Francisco on Tuesday, billing them as “always-on” AI agents that help with everything from dinner plans to website launches.
These agents are not just chatbots waiting for your next question. They are designed to keep working in the background, crawling the web and learning about your habits, then popping up with answers before you even ask. Think of them as persistent sidekicks that pull in information from connected services to make decisions on your behalf.
The rollout starts today for subscribers to ChatGPT’s Pro tier, which costs $100 a month. That means the feature lands first with OpenAI’s paying customers, who can start by controlling one Dot at a time. The company expects to eventually roll out the ability to control multiple agents simultaneously.
What Dots Actually Do
Dots are powered by OpenAI’s GPT-6 Astra model. The key difference from a standard chatbot is that Dots are always listening and acting. They do not wait for a prompt; they watch your calendar, your messages, and your connected apps, then act on what they see.
In a demo shared by OpenAI, a user’s Dot noticed that their calendar showed they were working through dinner. The Dot responded by messaging two GrubHub food delivery options, complete with pricing for each. The user followed up with their food pick and when they wanted it ordered.
Another example involved a user collaborating with their Dot to launch a new website. The agent handled the setup while the user directed the process, showing how Dots can take on project management alongside personal errands.
How Dots Connect With You
Dots can be reached through ChatGPT, Slack, and Microsoft Teams, with your context shared across those modes. That means a reminder set in Slack can trigger an action in ChatGPT, and a note in Microsoft Teams can surface in your daily digest.
Pro users can also join a waitlist to text with their Dots in iMessage or RCS messaging on Android.
The Always-On Design
The “always-on” label is the centerpiece of OpenAI’s pitch. Unlike a single-turn chatbot experience, Dots constantly crawl the web and attempt to complete whatever task you assign to them. They pull context from connected apps and learn more about your preferences over time.
This is a significant shift from the traditional chatbot model, where the conversation ends after each exchange. Dots keep the conversation open, which changes how you interact with them.
Approval Before Sensitive Actions
Dots require explicit approval before taking more sensitive actions on a user’s behalf, like installing software or changing a password. Users can enable a Custom Rules tool to set explicit boundaries that they do not want agents to cross, as well as tasks that need more direct permission to complete.
This is a safety feature, not a convenience one. It means Dots can be helpful without being intrusive, which is a hard balance for AI assistants to strike.
A Personal Agent Breakthrough
Personal agents are currently having a breakthrough moment. Early this year, tech enthusiasts in Silicon Valley latched onto tools like OpenClaw. More recently, Meta released its Muse agent, which reached a large user base and climbed to the top of smartphone app download charts.
OpenAI’s Dots arrive now as a direct answer to that success. The company is making a play for the same market.
The Waitlist Rollout
The rollout follows OpenAI’s usual pattern. New features often launch first for the higher subscription tier, then roll out to a larger subset of users over time. Subscribers to ChatGPT’s Pro tier get access today, while the waitlist for iMessage and RCS texting is available for Pro users.
Users start by controlling one Dot, but the company is expected to eventually roll out the ability to control multiple agents at the same time.
The Comparison to Meta’s Muse
| Feature | Dots | Muse |
|---|---|---|
| Model | Powered by GPT-6 Astra | Not specified by OpenAI |
| Availability | Pro tier rollout today | Reached a large user base |
| Connection | ChatGPT, Slack, Microsoft Teams | Not specified by OpenAI |
| Context sharing | Shared across modes | Not specified by OpenAI |
| Approval model | Explicit approval required | Not specified by OpenAI |
The table shows that OpenAI has built a distinct path for Dots. The comparison to Meta’s Muse is telling. OpenAI is not just building another chatbot; it is building a system that lives in your pocket and stays engaged with you throughout the day.
What Comes Next
OpenAI has put a lot of effort into making Dots feel like helpers rather than overlords. The Custom Rules tool gives users control over what the agent can and cannot do, which is a meaningful step toward trust.
The rollout to Pro subscribers is a smart move. It lets OpenAI test the waters with paying customers before expanding to everyone. If Dots work for the people who pay the most, the logic goes, they will work for everyone.
The comparison to Meta’s Muse shows that the field is crowded. But OpenAI has a history of launching features that stick, and Dots have the technical foundation to become a lasting part of how people manage their lives online.
Whether Dots can match the popularity of Muse remains to be seen. But OpenAI has made its move, and the agents are live for anyone willing to pay $100 a month for the privilege.
The rollout is live today, and the first batch of users will be testing it within hours.
Key Facts Box
- Rollout: Live today for ChatGPT Pro tier subscribers
- Cost: $100 per month
- Model: GPT-6 Astra
- Access: One Dot at a time, initially; multi-agent support expected later
- Connection: ChatGPT, Slack, Microsoft Teams
- Texting waitlist: iMessage and RCS on Android for Pro users
See the video the story is built around at WIRED.
Get the Notebook.
The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.