OpenAI lets ChatGPT Pro users in the US text their Dot agent

What OpenAI just added

OpenAI now lets you reach its ChatGPT Dot agent through ordinary text messages. As Yahoo Tech reported, the idea is that you can give Dot a job without first opening the ChatGPT app. An agent, in this context, is an AI that carries out tasks for you rather than only answering questions.

The feature works with Apple's iMessage and with RCS, the newer messaging standard on Android phones. You can send Dot written messages, photos and voice notes. Dot can also start the conversation, texting you when it has an update worth your attention.

OpenAI's official ChatGPT account announced it on October 28, 2026 with a short line: "By popular request, you can now text your dot." The post says you can hand Dot something to take off your plate or something to keep an eye on.

How texting Dot works

Setup starts inside ChatGPT, on the mobile app, the desktop app or the web version on a computer. Find your Dot's profile and choose the Texting option. You then confirm your phone number, and Dot sends back a contact card that you save to your phone. After that, you message it like any other contact.

iPhone users text from the number they verified. Android users first need RCS switched on in their messaging app.

Once connected, you can assign a task, ask Dot to watch something for you, or send a photo or voice note. It replies when there is something to report. Dot draws on your ChatGPT conversations and the channels you have connected to decide where an update should go.

You are not stuck with its choice. You can tell Dot to text you only when something is urgent and to keep routine updates inside ChatGPT. That setting matters because a proactive assistant can easily become a noisy one.

Approvals and privacy

Some tasks need your permission before Dot acts. In those cases it does not ask you to approve inside the text thread. It sends a link that opens the request in the ChatGPT app, where you can look at the details and decide.

On privacy, OpenAI makes two claims. Only you can text your Dot, and Dot can only send messages to you. The company also says that setting this up does not give Dot or OpenAI a view into the rest of your text conversations.

Those are the company's own statements, as relayed by Yahoo Tech. The report does not describe any independent check. Treat them as a starting point and read the settings yourself before you rely on the feature for anything sensitive.

Who can use it, and what is missing

This is a beta, and it is narrow. It is rolling out only to ChatGPT Pro subscribers in the US. People on other plans, or outside the country, have to wait for a wider release. The report does not say when that might happen.

The texting is strictly one-to-one. Group chats are not supported, and neither are phone calls or FaceTime. Calling your Dot from inside ChatGPT still works, so voice is not gone entirely. It just does not run through your phone's regular calling.

Our newsletter noted that the headline does not list what Dot can do by text. The report does not either, beyond handing off tasks and monitoring things. A sensible first test is a small errand where a mistake costs you little.

Why the messaging app is the point

Most people check their messages many times a day, and far fewer open an AI app on purpose. Putting an agent in the same thread as your friends and family removes a step. Short, casual instructions suit that setting.

The proactive part is the bigger shift. Until now, most chatbots waited for you to speak first. A Dot that texts you when something changes behaves more like a colleague who sends status notes. That is useful if you set the rules for when it speaks up, and annoying if you do not.

For readers outside the beta, the practical lesson still applies. Agents work best with clear, small, one-job requests, and the habit of writing them that way is worth building now.