OpenAI has launched Computer History, a new feature in ChatGPT's macOS app that records your clicks and keystrokes to help the AI understand your workflow and suggest automations.
Unlike Microsoft's Windows Recall, it does not capture screenshots, instead relying on event data.
The feature is opt-in and lets you exclude apps, websites, and private browsing from tracking.
What happened
OpenAI has rolled out Computer History, a new feature in ChatGPT's macOS desktop app that records your actions—clicks, keystrokes, and browsing—to build a timeline ChatGPT can reference when you make requests. The feature is opt-in, lets you exclude specific apps and websites, and automatically ignores incognito or private browser tabs.
Why it matters
Computer History turns your activity into training data so ChatGPT can suggest automations, complete half-finished tasks, and provide context about your work without requiring you to manually explain your situation each time. Unlike Microsoft's Windows Recall (which was controversial for taking screenshots), Computer History does not capture images, videos, or audio—only event data—giving you more granular control over what the AI learns about you.
What to watch
The feature is available on macOS now and is opt-in rather than opt-out. You can delete individual Computer History entries and exclude specific applications or websites from being tracked, allowing you to decide which parts of your workflow are fair game for AI analysis.
Ask the AI about this article →
Computer History represents OpenAI's answer to the challenge of giving ChatGPT meaningful context about a user's work without requiring constant manual explanation. By logging actions and building a timeline, the system lets ChatGPT reference what you were doing when you ask for help—looking up the last document you edited, checking if it was shared via Slack, or recapping your morning. This mirrors the intent behind Microsoft's Windows Recall feature, which faced public backlash over privacy concerns tied to its reliance on continuous screenshots. OpenAI's design choice to record events rather than images, videos, or audio is a deliberate privacy differentiation, positioning Computer History as a less intrusive alternative. The opt-in architecture and granular exclusion controls further suggest OpenAI is trying to balance the utility of activity tracking with user agency—allowing those who want AI that understands their workflow to opt in, while giving those who do so the ability to carve out zones of privacy.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
Ask AI anything about this article. Q&As are published on this page for other readers too.
Ora, a platform that tests how well AI agents can navigate and transact on live websites, ran side-by-side ben…
Panasonic Avionics Corporation built an agentic AI system on AWS using Amazon Bedrock, Amazon SageMaker, and A…
AWS introduced the Agentic Data Operations Platform (ADOP), a reference architecture built on Amazon Bedrock t…
Amazon Web Services introduced AgentCore Gateway, a new capability of Amazon Bedrock AgentCore that centralize…
Amazon Bedrock now supports query-aware compression, a pattern that filters retrieved context through a smalle…
Anthropic plans to "match or beat" the size of SpaceX's $75 billion IPO (or $86.2 billion including the over-a…
