ChatGPT Adds Sketch Tool to Convert Doodles into AI Images

OpenAI unveiled a new “Sketch” feature in ChatGPT Images 2.5 that lets users generate realistic images from simple doodles. By typing @Sketch into the chat, a drawing window appears, allowing users to sketch an idea and then describe how they want it rendered. In a brief test, a user drew a rough cat with a computer mouse and instructed ChatGPT to produce a realistic photo; the system successfully produced a detailed image that matched the description.

The update also adds the ability to comment directly on parts of generated images, giving users finer control over revisions and feedback. The feature is part of OpenAI’s broader push to make image creation more intuitive and interactive, building on the capabilities introduced in the earlier ChatGPT Images release. The Verge reports that the Sketch tool is now live for all ChatGPT Plus and Enterprise subscribers, expanding the platform’s creative toolkit.