Google Labs
Google's AI and product experiments
Start new thread

Google Whisk - Generate images using images as prompts, not text prompts

by•
Google's new AI tool, Whisk, allows you to generate images by using other images as prompts. Simply drag in images for the subject, scene, and style, and remix them to create unique, personalized visuals without needing long text prompts.

Add a comment

Replies

Best
Hunter
📌
Whisk is a new generative AI tool that lets you create images by inputting images, not text. You can drag in images for the subject, scene, and style, and then remix them to create something unique. Whisk uses Gemini to automatically write a detailed caption of your images, which is then fed into Imagen 3 to generate the final image. This process captures the essence of your subject, not an exact replica, allowing you to easily remix your subjects, scenes, and styles in novel ways. Try out Whisk in the US at labs.google/whisk and share how you like it. :)
At least, I joined waitlist as it is not available in more European countries. 🤦‍♀️
Hunter
you're right! Currently available in US and Canada, maybe they will roll out in other countries soon. :)
Hunter
ah, that’s smart! Haha 😁
Probably, but EU is so limited and restrictive about AI. Almost every feature comes here with delay because of policies.
Tried and the same result :)
Hunter
Ah yes, I heard they have EU AI Act to regulate artificial intelligence.
This looks very compelling — Google is on a tear lately!
Hunter
Haha yes, Google is on a roll — Imagen 3, VEO 2, Gemini 2.0, Whisk — they are pushing boundaries. Are you keeping up with the OpenAI updates too? Things are heating up!! 🔥 The competition is driving incredible innovation. :)
Makes tonnes of sense for this to be the next-step. I often really like the make sure that I use images as references when writing prompts, so this is going to be a game-changer for me no doubt. Google can't be stopped, they are killing the game!!
Hunter
I found it very innovative since not everyone is able to give detailed image prompts. :)
The "Studio Ghibli" style in the fish illustration at the end there is pretty cool. Hopefully the artists work that this is trained on gets a piece of the action 😅
Hunter
I love it too!! However, you have raised a good point, since it allows any images as inputs, not sure if artists and original creators get any credit.
Google's still killing it with user experience! 🎯 Let's see what other cool stuff they've got coming.
Google has really taken a different path from OpenAI. I hope everything goes well for them.
Makes tonnes of sense for this to be the next-step. I often really like the make sure that I use images as references when writing prompts, so this is going to be a game-changer for me no doubt. Google can't be stopped, they are killing the game!!
Whisk sounds like a creative powerhouse! The ability to generate unique visuals by using images as prompts—dragging in subjects, scenes, and styles for seamless remixing—is a brilliant approach. It’s a game-changer for personalized content creation without the hassle of crafting long text prompts. Excited to see the amazing art it will inspire! 🎨✨
This looks really promising! Unfortunately, it’s not available in my country, India, at the moment. Looking forward to the day I can give it a try. Hope to see it launched here soon!