ChatGPT can now turn your messy doodles into perfect images

OpenAI announced a new version of ChatGPT Images (called version 2.5) with a creative twist: you can now draw directly inside ChatGPT and have it transform your sketch into a finished image.

Here's how it works. You type @Sketch in your ChatGPT chat box, and a drawing window pops open. Then you draw whatever you want—even if you're not an artist. Once you've finished your rough doodle, you tell ChatGPT what you'd like it to create. The AI then takes your simple sketch and turns it into a detailed, polished image based on your instructions.

Why does this matter? Many people struggle with describing images in words. Maybe you have a vague idea in your head but can't quite put it into sentences. With Sketch, you can just draw it—badly is fine—and ChatGPT handles the rest. In testing, even a mouse-drawn cat doodle became a realistic photograph. This makes creative tools more accessible to everyone, not just people who can write detailed descriptions.

OpenAI also added another small but useful feature: you can now leave comments directly on parts of images you've created, making it easier to give feedback or ask for changes to specific areas.