en.Wedoany.com Reported - OpenAI has introduced the Record & Replay feature for Codex, allowing users to demonstrate a workflow to the agent once on a Mac. The agent then converts it into a reusable skill, automatically executing the task in the future without the need to write instructions in prompts.

During recording, Codex captures user actions and window content. When recording stops, the language model analyzes the sequence of steps and generates a SKILL.md file—a natural language instruction document describing the entire workflow. The completed skill can run in Computer Use mode, browser operations, and on top of connected plugins. In OpenAI's demonstration, Codex learned to upload videos to YouTube through this method, covering steps such as selecting files, filling in titles and descriptions, setting thumbnails, and loading subtitles.

Unlike classic RPA tools, Codex does not memorize click coordinates but generates text instructions, which are then interpreted by the reasoning model. Therefore, the skill can be applied to new data and the latest state of applications, rather than repeatedly executing hardcoded action chains that fail when the interface changes slightly. Recorded skills can be shared with teams, serving as ready-made automation solutions.

Record & Replay is currently only supported on macOS and is available to ChatGPT Plus, Pro, Business, Enterprise, and Edu subscribers, but not in the European Economic Area, the United Kingdom, or Switzerland. This follows OpenAI's classic release pattern, with Windows and other regions to be added later. Users need to enable Computer Use mode to use this feature.

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com