Mixup is a new, Mad Libs-style app for creating AI images from photos, text and doodles

A talented team of former Google employees, known for their innovative 3D design app Rooms, has introduced an exciting new project called Mixup. This iOS-exclusive application serves as a whimsical AI photo editor, allowing users to craft unique images through engaging prompts reminiscent of Mad Libs. With Mixup, users can transform their photos, sketches, and text into imaginative AI-generated visuals. For example, users can take their doodled sketches and reimagine them as stunning Renaissance artwork or visualize their pets in quirky Halloween outfits. The app even permits users to experiment with their selfies, exploring new hairstyles or even transforming a friend into a humorous caricature. Built upon Google’s advanced Nano Banana model, Mixup innovatively utilizes a recipe format that turns image generation into a social game. Jason Toff, CEO of Things, Inc., shared insights on the technology, highlighting how Nano Banana uniquely preserves the essence of the original image while generating new interpretations. Mixup’s standout feature is its shareable user-generated prompts. Toff explained that many existing generative AI tools often leave users scratching their heads, wondering what to create. Mixup alleviates this dilemma by allowing users to fill in the blanks of prompts that have already yielded successful results. Upon creating a prompt, users can choose to share their creation on a public feed or keep it private. Other users can then view these images and experiment with the same recipe, generating their own versions using personal photos or in-app doodles. This functionality also addresses the unpredictability often associated with generative AI by linking the resulting images directly with their prompts, enhancing user control over the creative process. Mixup also tackles the common frustration with generative AI outputs, which can feel random and uncontrollable. By allowing users to see both the original photo and the generated image alongside the prompt, the app instills a sense of predictability in the creative process. Users can also toggle a feature to view the transformation process, comparing the original and the AI-enhanced images. In addition, similar to OpenAI’s Sora app, Mixup lets users upload their own images for use in creations. This feature, termed “mixables,” allows friends to generate AI images using each other’s likenesses, fostering a community of creativity. While users can choose to keep their images private, the potential for a new creator class on the platform exists, enticing users to explore their creativity in unconventional ways. To address moderation concerns, Mixup incorporates OpenAI technology while also utilizing Google’s built-in safeguards to filter inappropriate content. Currently optimized for iOS 26, the app is compatible with iOS 18 and newer versions. As the app gains traction, the possibility of a web version or an Android counterpart may arise. Mixup provides users with 100 credits, valued at $4, to start. Each image generation costs approximately 4 cents, and once credits are depleted, users can opt for subscription tiers offering 100, 250, or 500 credits monthly. The app is set to launch globally on November 20 at midnight on the App Store, though entry will require an invitation. Users can pre-order Mixup now to be among the first to experience its creative possibilities.