Things Inc., the company founded by former Googlers known for their 3D design app Rooms, has unveiled its latest innovation: Mixup. This playful AI photo editor, exclusively available on iOS, empowers users to generate unique AI images from their photos, text, or even simple doodles. Mixup introduces a novel approach using "recipes" – Mad Libs-style, fill-in-the-blank prompts that guide the AI in transforming creative input.
Users can leverage Mixup to achieve a wide array of creative outcomes. Imagine transforming a quick sketch into a detailed Renaissance painting, or reimagining a beloved pet in a humorous Halloween costume. The app also allows for personal experimentation, such as previewing new hairstyles on a selfie or even generating whimsical, bizarre scenarios like envisioning a friend as "Italian brainrot."
Mixup is built upon Google's Nano Banana model, a technology praised by Things Inc. founder and CEO Jason Toff for its ability to convincingly integrate user images without uncanny results. Toff, an industry veteran with experience at Google, Meta, and Twitter, highlights how Mixup's "recipe" format redefines interaction with generative AI, transforming a powerful tool into an engaging, almost game-like experience.
A key feature enhancing Mixup's appeal is the shareability of its user-generated AI "recipes." Toff explains this innovation addresses a common frustration with generative AI: the blank canvas problem. "Generative AI is so powerful, and yet most of the time you go to these tools and it's like, here's your text box — come up with something creative. And what do you write?" he noted. Mixup mitigates this by allowing users to browse successful creations and simply "fill in the blanks" of existing recipes, fostering creativity without the initial hurdle of prompt engineering.
Upon creating a new image, users can either download it for personal use or publish it to a public feed, complete with the "recipe" used. This public feed allows others to "Try recipe" with their own inputs, promoting a collaborative and experimental environment. This design also tackles what Toff calls the "slot machine problem" of generative AI, where outputs can feel random and uncontrollable. By showcasing the original image alongside its recipe, Mixup provides users with a clear expectation of the AI's potential output, offering a sense of control and predictability. An optional toggle also allows viewing of before-and-after images.
Mirroring functionalities seen in tools like OpenAI's AI video app Sora, Mixup allows users to upload their own photos for AI image generation. This enables a unique feature called "mixables," where individuals you follow within the app can create AI images using your likeness. While envisioned as a fun feature for friend groups, it also opens the door for a new class of creators, provided they are comfortable with their image being transformed in potentially unconventional ways. Users retain control, with the option to avoid uploading photos or following others if they prefer not to have their likeness used.
To address moderation concerns, Mixup integrates OpenAI technology and relies on Google's inherent controls within its image model to filter inappropriate content. The app is optimized for iOS 26, supporting iOS 18 and newer versions, with potential plans for web and Android expansion. Upon launch, users receive 100 free credits (equivalent to $4), with each image costing approximately 4 cents. Subscription tiers are available for additional credits. Mixup officially launches globally on the App Store at midnight on November 20, requiring an invite for access, and is currently available for pre-order.
