MorseSign inStart freeOpen Morse

Are ChatGPT dots private? What your dot reads, keeps, and trains on

The Morse teamPublished 9 min read

ChatGPT dots are private from other people, but as of November 2026 they aren't private from OpenAI. A dot reads your ChatGPT memories and the apps you connect, can review that information and save notes when you haven't asked anything, and keeps what it learns "for as long as you keep your dot" (OpenAI). On Pro, its conversations, actions, and connected-app data can train OpenAI's models unless you turn off Improve the model for everyone, and you can't see or delete its memories one at a time.

We read OpenAI's dots privacy FAQ, its getting-started and workspace pages for dots, the ChatGPT release notes, and its pages on data controls, chat retention, and memory, plus the US privacy policy, on November 15, 2026. OpenAI deserves credit for the dots privacy FAQ. It says in plain words what a dot doesn't keep, what deleting it leaves behind, and that people may review a dot's activity even with training off.

ChatGPT dot privacy at a glance

Setting or featureDefaultHow long it's keptSource
Dot context and memoriesShared with ChatGPT Memory in both directions"For as long as you keep your dot"; deleting the dot starts deletionOpenAI
Proactive researchPart of every dot; Enterprise admins can't turn it off separatelySaved as the dot's private notesOpenAI
Connected appsOff until you connect one; permissions shared with ChatGPT, Work, and CodexDisconnecting leaves what the dot already learnedOpenAI
Model training (Pro)Improve the model for everyone on (OpenAI)De-identified copies can outlast deletionOpenAI
Texting (beta)Off; you verify a phone numberThe texting provider keeps no content after delivery; OpenAI stores texts with the dot's conversationsOpenAI
Chats, files, and Codex threads the dot madeStored apart from the dotSurvive deleting the dot; deleted chats leave OpenAI's systems within 30 daysOpenAI

What is a ChatGPT dot, and who has one?

A dot is an always-on agent with its own cloud computer. OpenAI introduced dots on September 29, 2026, for eligible Pro and Business Premium users aged 18 and older, and Pro access excludes the European Economic Area, Switzerland, and the UK (OpenAI). In Enterprise, Edu, and Healthcare workspaces, an admin preview runs from November 4 to November 18, 2026, and starting November 18 "dots become available by default unless your admin opts out" (OpenAI).

A dot's background research resembles what ChatGPT Pulse did overnight, with far more reach. OpenAI retired Pulse in June 2026 (OpenAI), and we cover the scheduled tasks that replaced its daily briefings in are ChatGPT scheduled tasks private.

What can a ChatGPT dot see?

Your memories, your recent chats, and every app you've already connected to ChatGPT. OpenAI says a dot "can receive memories and recent conversation context from ChatGPT," and that conversations with the dot "can contribute to memory in ChatGPT" (OpenAI). Plugin permissions are "shared across dots, ChatGPT, ChatGPT Work, and Codex," so a Gmail or Slack connection you set up months ago for ChatGPT is already open to the dot.

Your own computer is off limits until you allow it. Access to your local computer "is optional and starts turned off," and you grant it from the ChatGPT desktop app on that machine (OpenAI). The camera, microphone, and screen need that connection plus the ChatGPT app's permission in your device settings.

Does your dot read your apps when you haven't asked?

Yes. OpenAI calls this proactive research. The dot "can read information from permitted connected sources and save private notes" before you make a request (OpenAI). Its research tools can't send messages to other people, change content through plugins, or control a browser or computer. Enterprise admins "cannot turn off proactive research separately" (OpenAI).

What the Gmail permission itself allows is in is it safe to connect Gmail to AI.

What does a dot keep, and for how long?

Everything it has learned, until you delete the whole dot. OpenAI's privacy FAQ says a dot "can retain context from your conversations and plugins for as long as you keep your dot," and that "You currently cannot view, delete, or directly modify individual dot memories, including specific details that enter the dot's context from plugins" (OpenAI).

Some things it doesn't hold. The FAQ says the dot's context "does not retain credentials, images, or screenshots." Files it makes or uses elsewhere follow that place's rules, and deleting a Library file "does not by itself delete a copy already attached to your dot's conversation."

Turning off Memory in ChatGPT stops the sharing between ChatGPT and the dot, but "does not delete information that your dot has already received." The ChatGPT side has its own controls, covered in how to turn off ChatGPT memory. Purchases follow the same rule, and what a dot remembers from shopping is in ChatGPT Wallet privacy.

Does OpenAI train on what your dot does?

On personal plans, yes, unless you opt out. Data sharing "is enabled by default" on Free, Plus, and Pro (OpenAI). For dots, OpenAI says the Improve the model for everyone setting covers "actions dots take, work they delegate to other agents, automations you set up, and data from connected apps that is used to inform your conversations" (OpenAI). It says it removes personal identifiers "where possible."

The background research has its own line. OpenAI says "We don't train directly on proactive background research or its notes," but once the dot brings that research into a conversation or task, "that information may be used for training, depending on your settings." So an email the dot read in the background can reach training as soon as it shows up in a reply to you.

Business, Enterprise, and Edu workspaces aren't used for training by default (OpenAI). The rest of ChatGPT's training rules, including thumbs-up feedback, are in does ChatGPT train on your data.

Can people at OpenAI see your dot's activity?

Yes, in limited cases. Asked whether people associated with OpenAI can review a dot's activity with model improvement off, the FAQ answers "Yes. Human review may occur in limited circumstances, including safety-related cases" (OpenAI). Service providers for "hosting, cloud infrastructure, support, safety, and email delivery" may also process it, and OpenAI says content is encrypted in storage and in transit. Who those reviewers are is in is ChatGPT private.

On a work account, admins can review every personal dot in the workspace, with its creator, active or paused status, saved cost tier, and dot ID, and they can pause or delete it (OpenAI).

Are texts to your dot end-to-end encrypted?

No. OpenAI sends and receives a dot's texts through Linq, a third-party messaging provider, and says "the conversation is not end-to-end encrypted between you and your dot, even when you use iMessage or RCS" (OpenAI). Linq "doesn't retain their contents after delivery" but may keep timestamps and delivery status. OpenAI stores your texts with the dot's conversations.

Texting is a beta for a limited number of Pro users in the US, Canada, and Mexico, and you verify a phone number to turn it on. OpenAI says it gives no access to your other text conversations, other people can't text your dot, and the dot "can't participate in group texts." Reply STOP, or open the dot's profile and choose Texting, then Turn off texting. That disconnects your number but doesn't delete past conversations or memories, "either immediately or after 30 days."

Can a dot be tricked into leaking your data?

It can, and OpenAI says its protections "help reduce the risk of malicious instructions causing an unwanted action, but they do not eliminate it" (OpenAI). Before a dot sends an email, a check called Auto-review looks at the recipient and message. For supported sign-ins, your dot pauses while you type your password into a secure login form that the model doesn't see (OpenAI), but that doesn't cover passwords "you share separately in a chat or document, or through a plugin." How these attacks work is in what is prompt injection.

What each dot control removes, and what it leaves

Each row comes from OpenAI's getting-started page, privacy FAQ, and data controls page, read on November 15, 2026.

ControlWhat it stopsWhat it leaves
Pause the dotIts work until you resumeAll its information; separate Work, Codex, and scheduled tasks may continue
Disconnect an appNew access through that connectionWhat the dot already built into its context
Turn off ChatGPT MemoryMemory sharing between ChatGPT and the dotWhat the dot already received
Turn off textingTexts with your phone numberPast conversations and saved memories
Turn off Improve the model for everyoneTraining on the dot's conversations and workHuman review in limited cases
Delete the dotStarts deletion of its own contextChats, files, and Codex threads it made; your ChatGPT memories; separate scheduled tasks

To delete a dot, open its profile, select the more options menu (•••), then Delete, and confirm. OpenAI says your main conversation is archived, "Deletion can't be undone," and a new dot "does not restore the deleted dot's context." Then delete the chats it made. OpenAI schedules a deleted chat for permanent deletion "within 30 days," unless it was already de-identified or must be kept for security or legal reasons (OpenAI).

How Morse keeps a long-running project private

A dot can work while you're away because OpenAI holds your memories, your app connections, and everything the dot has learned on its servers, tied to your account, until you delete the dot itself.

Morse doesn't keep that copy. Your chat history is stored in your browser, encrypted, not on Morse's servers, and it locks when you sign out. Our servers don't keep your messages or log what you write. Morse doesn't train AI on your chats, sell your data, or show ads. Paste in only the part of the email or plan the question needs.

Pick the mode for the project:

  • Private, for the long-running plan you'd rather no host keep. New chats start here. Open-weight models like DeepSeek, Kimi, and GLM, plus xAI's Grok, answer with your account, email, and IP address stripped. Morse has zero data retention agreements with its inference, search, and other providers, so they've agreed to keep nothing, and where a provider confirms zero retention on each reply, Morse checks it and refuses a reply that doesn't.
  • E2EE, for the parts no one else should read. Your browser checks the secure enclave's hardware attestation before anything leaves your device, then encrypts the message so only that enclave can read it. Our servers relay ciphertext they can't read. It rests on the enclave hardware and attestation working as designed, and it has no web search or image generation.
  • Anonymous, to keep asking GPT without the provider knowing it's you. Our proxy removes your account, email, and IP address and sends the message under Morse's own account. The provider sees the words without knowing it's you and may keep them for a limited time to check for abuse, under its own policy, so leave out names and account numbers.

Because we don't store chats, we usually have very little to give authorities. Clearing your browser or losing the device deletes those chats, and Morse can't recover them, so keep finished work in your own files.

So is a ChatGPT dot private enough for you?

  • Your inbox and apps: connect only what the dot's goal needs, since it can read them without being asked and disconnecting doesn't erase what it learned.
  • Training: turn off Improve the model for everyone before you start. Morse doesn't train AI on your chats.
  • What it remembers: plan to delete the whole dot, then the chats it made, because single memories can't be removed.
  • Texts: skip texting for anything sensitive, since the messages aren't end-to-end encrypted. For anything no one should read, start an E2EE chat on Morse.

Send a first message from the Morse home page without an account, or start free with 15 messages a day. E2EE chats come with Pro at $19 a month.

Common questions

Can my dot read my other text messages?

No. OpenAI says turning on texting only lets you message your dot and doesn't give your dot or OpenAI access to your other text conversations. Other people can't text your dot, and it can't join group texts.

Does pausing my dot delete what it knows?

No. OpenAI says pausing doesn't delete the dot's information, and separate Work, Codex, and scheduled tasks may keep running until you stop them yourself.

Can my employer see my dot?

On an Enterprise, Edu, or Healthcare workspace, admins can list every personal dot with its creator, status, and dot ID, and can pause or delete it. From November 18, 2026, dots become available by default in those workspaces unless an admin opts out.

Can my dot see my passwords?

Not for supported sign-ins. OpenAI says your dot pauses while you type credentials into a secure login form that sends them to the browser without exposing them to the model, but that doesn't cover passwords you put in a chat, a document, or a plugin.

Ask what you actually want to ask.

80+ models, characters, and images, with every chat encrypted and kept on your device. Free to start.

Start a private chat

Read next