ChatGPT's New Computer History Tracks Browsing Like Recall

A New Kind of Memory for ChatGPT

OpenAI has introduced a feature that lets ChatGPT quietly observe what you do on your computer, building a searchable record of browser activity that the assistant can later reference. Coverage describes it as functioning much like Microsoft's controversial Windows Recall tool, but without the constant screenshotting — instead relying on tracked clicks, keystrokes, and browsing history to give ChatGPT and OpenAI's coding tool Codex more context about a user's work 1.

How It Works and What to Know Before Enabling It

The feature, referred to as ChatGPT Computer History, is opt-in rather than automatic, meaning users must deliberately turn it on before OpenAI begins logging their activity 3. Once enabled, it captures clicks and keystrokes across a browsing session, effectively giving the assistant a memory of what a person has been doing online so it can answer follow-up questions or pick up tasks with fuller context 3. Because the tool monitors granular input behavior rather than just page visits, it raises the same category of privacy questions that have dogged Recall since Microsoft first unveiled it, even though ChatGPT's approach avoids storing visual screenshots 13.

Part of a Broader Feature Race

The rollout comes as OpenAI works to close feature gaps with rivals. Anthropic's Claude has held an edge in several areas — including native iPhone app integration and dashboards that update themselves — that ChatGPT has only recently begun to match, and reporting suggests OpenAI is still catching up in some respects even as it pushes new capabilities like Computer History 2. This pattern reflects the intensifying competition among AI assistants, where companies are racing to make their tools feel more proactive and context-aware, whether through deeper browser integration, health-focused personalization apps that let users feed in medical data for tailored insights 5, or memory features that blur the line between assistant and operating-system utility.

Safety Concerns Overshadow Product Expansion

At the same time, OpenAI has been rolling out ChatGPT for Teens, adding safeguards such as scheduled breaks and quiet hours, as the company faces lawsuits and regulatory scrutiny tied to alleged harms suffered by younger users 4. Early impressions of that teen-focused effort have been sharply critical, with observers arguing the changes fall short of meaningfully protecting young people and calling the launch a stumble rather than a fix 6. Taken together, the coverage paints a picture of a company simultaneously pushing ChatGPT deeper into everyday computer use — tracking browsing, personalizing health guidance, and competing feature-for-feature with Claude — while struggling to reassure regulators, parents, and privacy advocates that these expansions are being deployed responsibly, particularly where teenagers and sensitive personal data are involved.