Are ChatGPT scheduled tasks private? What replaced Pulse, and what it reads overnight

The Morse teamPublished November 7, 20268 min read

ChatGPT scheduled tasks, which took over ChatGPT Pulse's daily briefings on June 17, 2026, are private from other people unless you share a task link, but not from OpenAI (OpenAI). As of November 7, 2026, a task can run in the background on a schedule you set, search the web, check connected apps such as Gmail, Slack, and GitHub, and read ChatGPT Health and ChatGPT Finances where you use them (OpenAI). What it reads from apps can train OpenAI's models while Improve the model for everyone is on (OpenAI).

We read OpenAI's release notes, its Pulse announcement, and its help pages on scheduled tasks, sharing, connected apps, Google app data, memory, data controls, chat retention, and ads on November 7, 2026. The Pulse help article itself now returns "That page doesn't exist." OpenAI deserves credit for its task and Google pages, which say in writing what pausing, deleting, and disconnecting each leave behind.

ChatGPT scheduled tasks privacy: the key facts

What happened to ChatGPT Pulse?

OpenAI shut it down. The June 17, 2026 release notes say "Pulse is being sunset as proactive updates move into scheduled tasks," and that Pro users would "continue to have access for 14 days." OpenAI's suggested substitute is to "ask ChatGPT to schedule a daily briefing based on your interests and past chats."

Pulse stayed a Pro feature for its whole life. It launched on September 25, 2025 as "an early preview" for Pro users on iOS and Android, reached the web on October 29, 2025, and the December 17, 2025 notes still said "Pulse is only available in ChatGPT Pro" (OpenAI). The announcement spoke of "the goal of making it available to everyone," and as of November 7, 2026 it still carries no note that Pulse is gone.

What Pulse read while you slept

The announcement said Pulse worked overnight. "Each night, it synthesizes information from your memory, chat history, and direct feedback," and you could add Gmail and Google Calendar, which were "off by default." The launch notes added that Pulse "requires saved memory and chat history to be turned on." Each morning's update was "available for that day only unless you save it as a chat or ask a follow-up question, which adds it to your conversation history."

What can a ChatGPT scheduled task read?

More than Pulse could, if you connect it. OpenAI's scheduled tasks page lists these sources:

• The web and your apps. "Monitoring tasks can search the web and check connected apps for changes," the June 17 notes say, and the task page adds that they "can use information from previous runs."
• Your inbox, as mail arrives. Event-triggered tasks run in ChatGPT Work and respond to "supported Gmail, Slack, or GitHub activity," such as a new email filtered "by sender or subject." They need Plus or higher.
• Your health and money. Tasks "can use information from ChatGPT Health or ChatGPT Finances," such as "a portfolio update after the market closes" or a summary of "recent spending." How Health handles that data is in is ChatGPT Health private.

Project files are the exception. "If you create a task in a project, it cannot access uploaded files or files stored in that project."

The Google permission is wider than reading. OpenAI's Google app page lists the Gmail scope ChatGPT requests as gmail.modify, with no read-only Gmail scope beside it, while Drive, Docs, and Contacts get read-only ones. Actions still need your say-so in most cases. OpenAI says "An action that sends a message or changes external data may require approval," and the task "pauses until you review it" (OpenAI). The exception is the Allow all actions permission, which OpenAI warns "carries elevated risk because supported actions may run without another confirmation" (OpenAI).

When do ChatGPT scheduled tasks run?

Whenever you schedule them, up to once an hour on paid plans. Free users get one run a day at most, in "flexible scheduling windows, such as morning, afternoon, or night." Event-triggered tasks can fire "up to 30 times per hour and 720 times per day."

A run that posts nothing has still run the model. OpenAI says a completed task "can use your plan's usage allowance or workspace credits even when it does not post an update," because "The task still performs model processing in the background." Tasks you stop using "may automatically pause after a period of inactivity."

Does OpenAI train on what your tasks read?

It can. Training is on by default for personal Free, Plus, and Pro accounts (OpenAI), and on those plans, plus Go, OpenAI "may use information accessed from apps to train our models" while Improve the model for everyone is on (OpenAI).

Google apps get a narrower rule. OpenAI says it doesn't train "on data directly from connected Google apps," except data you submit as feedback, data you paste or upload, and data "included in ChatGPT's response" (OpenAI). A morning briefing that sums up your inbox is a ChatGPT response. On our reading of OpenAI's wording, the email details it repeats fall under that third exception unless you turn training off.

Rating a task result widens training. OpenAI says that after a thumbs up or down, "the entire conversation associated with that feedback may be used to train OpenAI models," even with training off (OpenAI). The switch is under Settings, Data controls, then Improve the model for everyone. Everything else that reaches training is in does ChatGPT train on your data.

Can ChatGPT create a scheduled task you didn't ask for?

Yes, one. OpenAI says "Connecting a financial account to ChatGPT Finances may create a Weekly finances update task," and existing Finances users "may receive a one-time notification when the task is added." To stop it, open Scheduled, select Weekly finances update, and choose Pause or Delete (OpenAI).

How do I turn off ChatGPT scheduled tasks?

Open Scheduled in the sidebar, or go to Settings, Notifications, then Manage tasks, and pause or delete each task. To cut off an app, open Settings, then Plugins, select the app, and choose Disconnect (OpenAI). Each control stops one thing and leaves others, so we built this table from OpenAI's pages on November 7, 2026.

So a clean exit takes three steps: delete the task, delete its chat, and disconnect the app. How share links outlive their chats is covered in AI chat share links, and what turning off Memory leaves is in how to turn off ChatGPT memory.

How Morse keeps your history away from jobs that run without you

Pulse and the tasks that replaced it work because OpenAI keeps your chat history, memories, and app connections on its servers, so ChatGPT can open them on a schedule while you're asleep or away.

Morse doesn't keep that copy. Your chat history is stored in your browser, encrypted, and locked when you sign out. Our servers don't keep your messages or log what you write, so there's no history on our side for anything to read overnight. Morse doesn't train AI on your content, sell your data, or show ads, so there's no training switch to hunt for. Because we don't store chats, we usually have very little to give authorities.

Pick the mode for what you'd have handed a briefing:

• Private, the default, for the daily questions a briefing would answer. Open-weight models like DeepSeek, Kimi, and GLM, plus xAI's Grok, answer with your account, email, and IP address removed. Morse has zero data retention agreements with its inference, search, and other providers, so they keep nothing.
• E2EE, for the money and health questions you wouldn't connect an app for. Your browser checks the secure enclave's proof that it's genuine 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 runs open-weight models with no web search, and it rests on the enclave hardware and attestation working as designed.
• Anonymous, to ask GPT without OpenAI knowing it's you. Our proxy strips your account, email, and IP address and sends the message under Morse's own account. OpenAI 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 names out.

Clearing your browser, losing the device, or deleting your account deletes your chats, and we can't recover them, so save anything you need. Our privacy policy lists what the account does hold: sign-in details, plan, balance, and a usage ledger with no message content.

Try a chat from the Morse home page with no account, or sign up free with no card. End-to-end encrypted chats come with Pro, $19 a month.