The Mixup app arrives with the kind of playful energy that makes AI feel less like a complicated tool and more like a creative game you can pick up anytime. Built by the ex-Google team behind the 3D design platform Rooms, this iOS app turns generative AI into something familiar and fun. Instead of staring at an empty prompt box and wondering what to type, you simply fill in a recipe. It works almost like a Mad Lib, but for AI image creation.
The idea is simple. You choose a recipe, drop in a photo, a scribble, or a quick line of text, and the Mixup app transforms it into something completely new. You can turn a rough doodle into a classical painting. You can imagine your pet in a wild Halloween outfit. You can test different hairstyles using a selfie or turn your friend into something wonderfully ridiculous. The app reduces friction at every step, so you feel guided rather than overwhelmed.
The Mixup team built the app on top of Google’s Nano Banana model, and they found that this underlying technology could keep the essence of your original image without sliding into uncanny territory. Jason Toff, the founder and CEO of Things, Inc., says this was a breakthrough. He has worked on experimental products at companies like Google, Meta, and Twitter, so he knows the delicate balance between capability and creepiness. With Mixup, he believes they landed in the sweet spot, where the image looks convincing yet still feels fun.
What makes the Mixup app even more entertaining is how these recipes work. They are not hidden system prompts. Anyone can create one. Anyone can share one. And anyone can remix one. A recipe sits beside its finished photo on a public feed. If you like it, you simply tap to use it with your own photo or your own scribbles. The app even includes a simple drawing tool for people who want to jump in without taking a picture first.
Seeing a finished image next to its recipe solves one of the biggest frustrations in generative AI. Normally, you type a prompt, press the button, and hope for the best. Then you try again. And again. The randomness begins to feel like a slot machine, where you never know what will appear next. Mixup gives people a clearer sense of control by showing what worked before. You can see the inputs. You can see the outputs. This transparency helps people understand how the image is shaped and what they can expect from their own attempt.
The before and after toggle makes that even more satisfying. You can quickly see how much of the image was transformed or preserved. The creator can choose whether to leave this on, but when it is available, it helps users understand what the AI actually did.
The Mixup app also lets you upload your own photos and make them available to people you follow. This feature, called mixables, opens the door to collaborative creativity. Friends can make wild mashups of each other. Small creator communities can experiment together. Some people might even emerge as characters within the app’s ecosystem. It feels social without feeling like another noisy social network.
If you prefer privacy, you can simply avoid uploading your image or avoid following anyone. The controls respect user choice. The team also uses safety tools from OpenAI along with guardrails inside Google’s image model to block sexual or violent outputs. Mixup tries to keep the fun without drifting into unsafe territory.
The app launches as an iOS exclusive optimized for iOS 26, but it also works on devices running iOS 18 and above. With enough traction, a web version or Android release could follow. For now, though, the team wants to focus on the iPhone experience. It feels polished, fast, and easy to learn.
The credit system keeps things lightweight. Free users receive 100 credits, which equal around four dollars worth of image generation. Each image costs a little under four cents to produce. When your credits run out, you can subscribe to monthly plans that offer 100, 250, or 500 credits. The pricing is simple enough for casual use but flexible enough for people who want to create daily.
Mixup opens to the public on November 21, and early adopters can join with an invite code. TechCrunch readers can use TCHCRH as long as it is still active. The app is already available for preorder on the App Store worldwide.
What stands out about the Mixup app is how it reframes AI image generation. Many tools promise creativity but leave people staring at a blank prompt. Mixup flips that experience. You begin with something that already worked. You tweak it. You try again. You share it. You learn what others create. This loop feels much more like a playful collaboration than a technical task.
The app also taps into something deeper. People want AI to feel friendly, accessible, and expressive. They want to experiment without pressure. They want to be surprised in good ways, not confused or overwhelmed. Mixup understands this desire. It gives structure without limiting imagination. It gives clarity without taking away the magic. It gives people a way to participate even if they never considered themselves artistic or tech savvy.
With time, Mixup could grow into a community where recipes become a new form of creative expression. Someone might create a recipe for renaissance pets. Someone else might design one for surreal landscapes. Someone might focus on cartoon styles or holiday themes. These recipes become tiny creative seeds that anyone can plant and grow in their own direction.
The Mixup app arrives at the perfect moment, when people are looking for softer, simpler ways to use generative AI. It blends entertainment with experimentation. It keeps things light while still feeling powerful. And it invites people to play with AI in a more social and intuitive way. If the team can maintain this balance as the app grows, Mixup could easily become one of those rare tools that reach beyond tech circles and into everyday culture.
