Google Labs has released Whisk, its latest generative AI experiment, in the United States. Unlike traditional image generators, which mostly rely on text prompts, Whisk focuses on using images as the primary input method.
Users can either upload images directly into Whisk or generate them within the tool, specifying elements for the subject, scene, and style. The system allows users to mix and match these components and fine-tune the results with additional text prompts if needed.
Behind the scenes, Google's language model—likely the recently released Gemini 2.0 Flash—creates detailed descriptions of the input images automatically. These descriptions then feed into Google's latest image generation model, Imagen 3, which captures the essential features of the subject rather than creating exact copies.
Creative tool, not a perfect copier
Since Whisk only pulls out a few key elements from each source image, Google warns that the results might not match what users expect. The generated images could come out with different heights, weights, hairstyles, or skin tones than the originals. Google knows these details can make or break a project, so it lets users see and edit the text prompts that drive the image generation process.
Early testers, including artists and creative professionals, say Whisk feels more like a new kind of creative tool than a standard image editor. Google built it for quick visual brainstorming rather than pixel-perfect editing, letting users quickly generate and sort through dozens of options before saving their favorites.
According to The Verge's hands-on testing, while Whisk is enjoyable to use, you'll need to wait a few seconds for each new image to generate. These delays could be temporary, though, likely caused by high traffic as people rush to try the new tool.
Where to try Whisk
Right now, Google is only letting users in the United States test Whisk. If you're in the US, you can try it for free at labs.google/whisk and share your feedback. Users outside the US won't be able to access the tool.
Whisk lives in Google Labs, the company's testing ground for experimental AI projects. This is where Google tries out practical applications for its AI models like Gemini, Imagen, and Veo, including its latest video model, Veo 2.
While most projects stay in the experimental phase, a few graduate to become full products - like NotebookLM, the company's AI research assistant that recently made the jump to general release.
