ChatGPT Can Now Remember What You Do on Your Mac, Without Taking Screenshots
OpenAI has launched Computer History for ChatGPT on Mac, an opt-in feature that turns activity across apps and websites into memories ChatGPT and Codex can use.
The part worth watching is what happens when an AI assistant stops waiting for context and starts arriving with a memory of how your day unfolded.
➜ The real story: OpenAI is trying to make ChatGPT feel less like a chatbot you brief from scratch and more like a coworker that already knows where you left off, but the privacy tradeoff now lives much closer to the keyboard.
Computer History replaces the earlier Chronicle experiment, which relied on screenshots to reconstruct what someone had been doing.
This version takes a different route: it uses interaction events from macOS to understand activity, then turns that context into a searchable timeline and local memories.
The goal is simple enough: come back after lunch, ask what you were working on, and skip the ritual of explaining your entire morning to ChatGPT.
How ChatGPT Computer History Works on Mac
Computer History can collect events such as clicks, typing, keyboard shortcuts, switching between apps and other context macOS exposes through its accessibility system.
It doesn’t capture screenshots, screen recordings, microphone input or system audio, and private browsing activity is excluded.
You can decide which apps and websites contribute, pause collection from the Mac menu bar, and delete individual history items or clear recent activity.
That gives ChatGPT and Codex a running sense of what you’ve been doing without literally keeping a photo album of your screen.
ChatGPT Computer History Privacy: What OpenAI Collects and Stores
The privacy story is better than Chronicle in one obvious way, but it still deserves attention.
Temporary interaction-event files can remain on the Mac for up to 48 hours, while generated memory files stay locally until you delete them; OpenAI processes temporary event data on its servers to create those memories and says the event files aren’t retained afterward or used for training.
The local memories themselves are plain-text files and can contain sensitive information, which is why OpenAI recommends excluding health, financial and other private apps and pausing the feature during communications with other people unless they’ve agreed to it.
That concern fits into the much bigger conversation around Shadow AI security risks and sensitive company data, where convenience can quietly expand what an AI system is allowed to see.
Why ChatGPT Computer History Could Make AI Assistants Much More Useful
This is where Computer History starts to look less like a memory feature and more like infrastructure for AI agents.
ChatGPT could identify the document you were reviewing, summarize yesterday’s work, notice a repetitive workflow or suggest turning something you keep doing into a reusable skill or automation.
That deeper context also creates a bigger security surface because malicious instructions hidden inside websites or apps can potentially make their way into what ChatGPT or Codex sees, a problem that becomes increasingly important as OpenAI AI agents become more capable in cybersecurity and autonomous computer tasks.
Still, the direction is pretty clear: the next useful AI assistant may not be the one that gives the smartest answer, but the one that already understands enough context to know what you were trying to accomplish.
Computer History Pushes ChatGPT Closer to an Always-Available AI Assistant
Computer History is off by default and initially limited to ChatGPT Pro, Business and Enterprise users on macOS, with administrators controlling access in Business and Enterprise workspaces before individuals can opt in.
It also isn’t initially available in the European Economic Area, Switzerland or the United Kingdom. Put this beside OpenAI plans for a more persistent ChatGPT hardware experience and the pattern gets interesting: ChatGPT is steadily moving beyond the blank chat box toward software and devices that understand what’s happening around you.
If OpenAI can make that context genuinely useful while keeping the controls understandable, Computer History could be one of those quiet features that changes how people expect AI assistants to work.
FAQs
Quick answers to the biggest questions about ChatGPT Computer History, Mac privacy and how OpenAI new memory feature works.
What Is ChatGPT Computer History?
ChatGPT Computer History is an opt-in Mac feature that turns activity across approved apps and websites into a timeline and memories that ChatGPT and Codex can use.
Does ChatGPT Computer History Record Screenshots?
No. Computer History records interaction events such as typing, clicks, keyboard shortcuts and app switching, but it doesn’t capture screenshots, screen recordings, microphone input or system audio.
Who Can Use ChatGPT Computer History on Mac?
Computer History is initially available to ChatGPT Pro, Business and Enterprise users on macOS. Business and Enterprise administrators must enable access before individual users can choose to turn it on.
