NanoAPI
One API for every creative model.
Whisk AI is an online creative workspace for AI image generation, reference remixing and image-to-video creation. Generate images from text, remix up to three reference images for subject, scene and style, or turn a still image into a short AI video. New users get 10 free signup credits; paid plans start at $9.90/mo. Whisk AI is an independent creative workspace and is not an official Google Labs product.
Whisk AI was an experimental artificial intelligence image-generation tool developed by Google Labs. Launched in December 2024, it introduced a visual approach to AI image creation by allowing users to generate and remix images using reference pictures rather than relying entirely on lengthy text prompts.
The platform was designed for artists, designers, content creators, marketers, and other creative professionals who wanted to explore visual ideas quickly. Instead of requiring detailed descriptions of every element, Whisk allowed users to upload images representing three main creative components: subject, scene, and style.
Whisk used Google’s Gemini models to analyse uploaded reference images and automatically generate detailed descriptions of their visual characteristics. These descriptions were then processed by Google’s Imagen image-generation model to produce a new image combining the selected elements.
Subject: Users could upload an image representing the main object, character, person, or element they wanted in their creation.
Scene: This input allowed users to provide a reference for the environment, background, or setting of the generated image.
Style: Users could select an image representing the artistic direction, visual appearance, colours, or overall aesthetic they wanted the final result to follow.
By combining these three inputs, Whisk made it possible to experiment with different visual compositions without writing complicated prompts. Users could also add text instructions to refine their results and adjust specific creative details.
One of Whisk’s defining features was its emphasis on visual exploration rather than precise image editing. Google’s approach focused on capturing the essence of uploaded reference images instead of reproducing them exactly. This allowed users to experiment with different interpretations of their ideas, although generated results could differ from the original references.
Whisk was particularly useful for creative brainstorming, concept development, character exploration, visual storytelling, product inspiration, and digital artwork. Content creators could experiment with different backgrounds, artistic styles, and subject combinations when developing visual concepts.
The platform also introduced Whisk Animate, which extended its creative capabilities by allowing users to transform generated images into short animated video clips using Google’s video-generation technology.
In March 2026, Google announced that Whisk’s best capabilities would move into Flow, its unified AI-powered image and video creation platform. Whisk was officially discontinued as a standalone service on April 30, 2026.
Although the original platform is no longer available independently, its subject, scene, and style remixing concept continues within Google’s broader creative AI ecosystem.
Whisk was offered as a free Google Labs experiment, with access and generation limits varying according to availability, location, and Google’s AI credit arrangements.
Free access: Available to eligible users during its standalone experimental period.
AI credits: Certain advanced generation features operated through Google’s AI credits system.
Paid access: Higher generation limits were associated with eligible Google AI subscription plans.
Enterprise: No separate standalone Whisk enterprise subscription was offered.
Whisk offered an innovative approach to AI image generation by making reference images the foundation of the creative process. Its subject, scene, and style workflow made visual experimentation accessible to users who preferred working with images rather than writing detailed prompts.
The integration of Gemini’s image-understanding capabilities with Imagen’s image-generation technology was one of its major strengths. It simplified the process of combining different visual references and exploring creative possibilities.
However, Whisk was not designed for pixel-perfect image editing. Generated results could differ from the reference images, and users sometimes needed additional text instructions to achieve their intended outcomes. Its experimental availability and eventual discontinuation also limited its long-term standalone usefulness.
Overall, Whisk was an innovative AI image-generation experiment that introduced an accessible approach to visual prompting. Its creative features remain relevant through Google’s transition to Flow.
One API for every creative model.
AI photo editor and image generator you can direct in plain language
Seedance 2.5 Flash AI video generator