How Whisk Works

Whisk allows users to merge three distinct images: one for the subject, one for the scene, and one for the style. This innovative approach generates brand-new images at the click of a button, offering incredible flexibility in customizing creations.

Whisk operates on the Imagen 3 model, a powerful image generator developed by Google. The process is simple yet highly creative. For instance, a user can select a photo of themselves as the subject, a futuristic landscape as the scene, and an anime style for the visual aspect. The generator then combines these three elements to create a new image.

This “visual remix” method enables users to explore unexpected combinations and bring fantastic images to life, all using their own photos.

Automated Yet Customizable Results

One interesting feature of Whisk is its ability to generate automatic captions. After selecting the images, the system creates a detailed description that guides the image generation process by Imagen 3.

For example, you could enter a prompt like “The subject rides a flying bike” to specify your creation further. However, Google notes that the results may vary. The generated subject can differ slightly in terms of size, weight, hairstyle, or complexion. Users can adjust these results by modifying the underlying prompts at any time.

Currently, Whisk is an experience available only to users based in the United States. Interested individuals can access it via labs.google/whisk. Google warns that, being a developing experiment, the results are not always perfect. Despite these limitations, Whisk offers an exciting glimpse into the future of visual creation.