{"id":2790,"date":"2026-08-14T08:16:06","date_gmt":"2026-08-14T08:16:06","guid":{"rendered":"https:\/\/dmarketertayeeb.com\/blog\/chatgpt-computer-history-codex-privacy\/"},"modified":"2026-08-14T08:16:53","modified_gmt":"2026-08-14T08:16:53","slug":"chatgpt-computer-history-codex-privacy","status":"publish","type":"post","link":"https:\/\/dmarketertayeeb.com\/blog\/chatgpt-computer-history-codex-privacy\/","title":{"rendered":"ChatGPT Computer History: What It Records and Who Should Use It"},"content":{"rendered":"\n<p><strong>OpenAI&#8217;s Computer History is an opt-in macOS feature that turns selected app and website activity into a timeline and local memory files that ChatGPT and Codex can use.<\/strong> It launched on 13 August 2026 for ChatGPT Pro, Business and Enterprise users outside the EEA, Switzerland and the UK. It does not record screenshots or audio, but it can capture clicks, typing, shortcuts, app switches and accessibility context\u2014so the privacy and security trade-off remains substantial.<\/p>\n\n\n<div class=\"dmt-fact-box\">\n\n<h2 class=\"wp-block-heading\">Computer History at a glance<\/h2>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Announcement:<\/strong> 13 August 2026<\/li>\n<li><strong>Status:<\/strong> Opt-in feature, off by default<\/li>\n<li><strong>Plans:<\/strong> ChatGPT Pro, Business and Enterprise<\/li>\n<li><strong>Platform:<\/strong> ChatGPT desktop app on macOS<\/li>\n<li><strong>Excluded regions:<\/strong> EEA, Switzerland and United Kingdom<\/li>\n<li><strong>Requirements:<\/strong> Memories enabled; workspace admin approval for Business and Enterprise<\/li>\n<li><strong>Not supported:<\/strong> API-key authentication and Amazon Bedrock<\/li>\n<li><strong>Price:<\/strong> No separate price disclosed; the feature uses model tokens<\/li>\n<\/ul>\n\n<\/div>\n\n\n<h2 class=\"wp-block-heading\">What Computer History records<\/h2>\n\n\n\n<p>Computer History creates an interaction-event stream from apps and websites you allow. Events can include clicks, text entry, keyboard shortcuts, application switching and context exposed by macOS accessibility APIs. OpenAI periodically converts that event stream into text summaries and local Markdown memories.<\/p>\n\n\n\n<p>It does not capture screenshots, screen recordings, microphone input or system audio. Private-mode browser activity is excluded. That makes it materially different from Chronicle, OpenAI&#8217;s earlier screenshot-based research preview. Computer History replaces Chronicle, but OpenAI describes it as a rebuilt system rather than a renamed version.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">What it can help you do<\/h2>\n\n\n\n<p>The feature is designed for questions that are difficult to answer from one chat alone: what you were doing before a break, which document you viewed, what tasks you worked on yesterday, or how a repeated workflow unfolds across several tools.<\/p>\n\n\n\n<p>When the original Slack thread, Google Doc, file or web page is available, the history can help identify the source and ChatGPT or Codex can then read that source directly. A timeline item can also suggest turning a repeated process into a skill or automation. The suggestion is not an automatic deployment; the user reviews it first.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Storage, retention and deletion<\/h2>\n\n\n\n<p>Temporary event files can remain on the Mac for up to 48 hours while the service creates summaries. OpenAI says those event files are processed server-side, are not retained there after processing unless legally required, and are not used for model training. The resulting memory files persist locally until the user deletes them.<\/p>\n\n\n\n<p>Those local memories are plain Markdown and are not encrypted by Computer History. Other software running as the same macOS user may be able to read them. FileVault and normal endpoint controls still matter, but they do not turn a broadly captured work history into a suitable secrets vault.<\/p>\n\n\n\n<p>Users can inspect or delete individual history items, clear recent windows of activity, clear all Computer History, pause collection from the menu bar or turn the feature off. Changing an allow or block rule affects future capture; it does not erase older history automatically.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">The practical privacy risk<\/h2>\n\n\n\n<p>&#8220;No screenshots&#8221; is useful, but it should not be confused with &#8220;no sensitive data.&#8221; Typed text, window context and accessibility information can still reveal client names, internal URLs, drafts, customer records or private conversations. OpenAI explicitly advises disabling Computer History during communications with other people unless they have given prior express consent.<\/p>\n\n\n\n<p>The safest rollout is include-only. Start with a code editor, terminal, project-management tool and selected documentation sites. Exclude password managers, finance and health services, private messaging, personal email and any system containing customer or credential data unless there is a documented need.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Prompt injection becomes a history problem<\/h2>\n\n\n\n<p>A malicious webpage or document can contain instructions aimed at an AI agent. If that content becomes part of a later memory, the attack may reappear after the user has left the original page. OpenAI therefore warns about prompt injection from captured apps and websites.<\/p>\n\n\n\n<p>For teams, the controls should include narrow source selection, permission profiles, sandboxing and explicit approval for consequential actions. A memory can help an agent find context; it should not silently widen the agent&#8217;s authority.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Useful marketing and operations workflows<\/h2>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Daily work recap:<\/strong> reconstruct the documents, dashboards and tasks used during a reporting cycle.<\/li>\n<li><strong>Campaign hand-off:<\/strong> identify the brief, approval thread and latest asset without knowing every file name.<\/li>\n<li><strong>SEO research continuity:<\/strong> recover the search result, competitor page and draft that informed an unfinished analysis.<\/li>\n<li><strong>Process documentation:<\/strong> turn a repeated publication or QA sequence into a reviewed skill.<\/li>\n<li><strong>Stand-up preparation:<\/strong> summarise recent work, then verify each item against the original source before sharing.<\/li>\n<\/ul>\n\n\n\n<p>Do not use Computer History as an automatic employee-surveillance system or as proof that a task was completed. The timeline is AI-generated context and still needs source verification.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Early community reaction<\/h2>\n\n\n\n<p>Discussion in the first 24 hours is limited. The clearest pattern is a split between continuity benefits and discomfort with continuous activity capture. Some early posts emphasise how useful a searchable workday could be; others focus on local plaintext memories, consent and prompt-injection exposure. The sample is too small and too new to represent users generally.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">DMT verdict<\/h2>\n\n\n\n<p>Individual Pro users with a well-secured Mac and a narrow set of work apps can test Computer History cautiously. Business and Enterprise teams should pilot it with include-only rules, a written data classification, consent guidance and a deletion routine. Anyone handling credentials, regulated information or sensitive communications should wait until the capture boundary and endpoint controls are demonstrably adequate.<\/p>\n\n\n\n<p>Computer History can feed the kind of governed workflows described in DMT&#8217;s <a href=\"https:\/\/dmarketertayeeb.com\/blog\/openai-codex-marketers-plugins-sites-workflows\/\">ChatGPT Work and Codex guide<\/a> and <a href=\"https:\/\/dmarketertayeeb.com\/blog\/chatgpt-work-scheduled-tasks-marketers\/\">scheduled-tasks operating guide<\/a>, but it does not replace their permission and review gates.<\/p>\n\n\n\n<p>Also see the <a href=\"https:\/\/dmarketertayeeb.com\/blog\/ai-agent-harness-context-compaction\/\">agent-harness controls<\/a>, <a href=\"https:\/\/dmarketertayeeb.com\/blog\/gpt-5-6-sol-codex-usage-272k-context-pricing\/\">Codex usage and context guide<\/a> and <a href=\"https:\/\/dmarketertayeeb.com\/blog\/chatgpt-for-digital-marketing-guide\/\">ChatGPT marketing guide<\/a> for adjacent workflows.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Sources<\/h2>\n\n\n\n<ul class=\"wp-block-list\">\n<li><a href=\"https:\/\/learn.chatgpt.com\/docs\/customization\/computer-history\">OpenAI documentation: Computer History<\/a><\/li>\n<li><a href=\"https:\/\/learn.chatgpt.com\/docs\/changelog\">OpenAI ChatGPT and Codex changelog<\/a><\/li>\n<li><a href=\"https:\/\/www.reddit.com\/r\/CodexAutomation\/comments\/1vnzmkf\/chatgpt_computer_history_gives_codex_a_searchable\/\">Reddit discussion: searchable work history and security trade-offs<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>Computer History is an opt-in macOS feature that turns selected activity into ChatGPT and Codex memories. See what it records, stores and risks.<\/p>\n","protected":false},"author":1,"featured_media":2791,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[180],"tags":[],"class_list":["post-2790","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-news","has-featured-image"],"_links":{"self":[{"href":"https:\/\/dmarketertayeeb.com\/blog\/wp-json\/wp\/v2\/posts\/2790","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/dmarketertayeeb.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/dmarketertayeeb.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/dmarketertayeeb.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/dmarketertayeeb.com\/blog\/wp-json\/wp\/v2\/comments?post=2790"}],"version-history":[{"count":1,"href":"https:\/\/dmarketertayeeb.com\/blog\/wp-json\/wp\/v2\/posts\/2790\/revisions"}],"predecessor-version":[{"id":2792,"href":"https:\/\/dmarketertayeeb.com\/blog\/wp-json\/wp\/v2\/posts\/2790\/revisions\/2792"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/dmarketertayeeb.com\/blog\/wp-json\/wp\/v2\/media\/2791"}],"wp:attachment":[{"href":"https:\/\/dmarketertayeeb.com\/blog\/wp-json\/wp\/v2\/media?parent=2790"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/dmarketertayeeb.com\/blog\/wp-json\/wp\/v2\/categories?post=2790"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/dmarketertayeeb.com\/blog\/wp-json\/wp\/v2\/tags?post=2790"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}