A team of ex-Google engineers just dropped Mixup, an iOS app that's reimagining AI photo editing as a social game. Built on Google's Nano Banana model, the app lets users create and share "recipes" - Mad Libs-style prompts that transform photos, sketches, and text into AI-generated images. Instead of staring at a blank prompt box, users can browse successful recipes from others and just fill in the blanks with their own content.
The timing couldn't be better for a consumer-friendly AI photo app. While Meta and OpenAI duke it out over enterprise AI dominance, Things Inc. just quietly launched something that could actually get your mom using generative AI. Mixup, which hits the App Store at midnight November 20, transforms the intimidating world of AI image prompts into something that feels more like playing Mad Libs with friends.
"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?" Things Inc. CEO Jason Toff told TechCrunch, pinpointing exactly why most AI tools remain stuck in the early adopter phase. The former Google and Meta product manager built Mixup to solve what his team calls the "slot machine problem" - that unpredictable, frustrating experience of never knowing what you'll get from an AI prompt.
The app's secret weapon is Google's Nano Banana model, which Toff says "could take your image and maintain it in a convincing way that wasn't creepy." But it's the social layer that makes Mixup feel revolutionary. Users can browse a public feed of "recipes" - successful AI prompts paired with their results - then tap "Try recipe" to apply that same transformation to their own photos, sketches, or text.
This isn't just another AI photo filter. The recipes range from turning doodles into Renaissance paintings to reimagining pets in Halloween costumes. One particularly viral early example transforms friends into "Italian brainrot" - because apparently that's where we are as a society now. The key insight: seeing both the prompt and the result removes the guesswork that makes most AI tools feel like gambling.
The economics tell an interesting story about where consumer AI is headed. At roughly 4 cents per image generation, Mixup sits in the sweet spot between free tools that compromise on quality and premium services that price out casual users. Free accounts get 100 credits (worth $4), with paid tiers offering 100, 250, or 500 monthly credits. It's a model that suggests AI content creation is becoming a mainstream behavior worth monetizing directly.
But the real innovation might be the "mixables" feature. Users can upload photos of themselves, making them available to anyone they follow for AI transformations. Things Inc. envisions friend groups following each other to create bizarre mashups, though they acknowledge a creator economy could emerge around people comfortable having their likeness remixed.
The moderation challenge looms large. Mixup combines OpenAI technology for content filtering with Google's built-in Nano Banana controls to restrict sexual content and violence. But as any social platform learns, user-generated AI content creates moderation nightmares that traditional photo apps never faced.
Timing-wise, Mixup arrives as the AI consumer app market heats up. OpenAI's Sora video generator is making waves in early testing, while Meta pushes AI features deeper into Instagram and Facebook. The iOS-only launch (with potential web and Android versions later) suggests Things Inc. is betting on quality over reach - smart positioning when competing against big tech's infinite resources.
The invite-only launch creates artificial scarcity that could drive early adoption, assuming the product delivers on its promise to make AI creation feel less like work and more like play. For an industry obsessed with prompt engineering and technical mastery, Mixup's "just fill in the blanks" approach feels both obvious and overdue.
Mixup represents exactly the kind of consumer AI breakthrough the industry needs - taking powerful but intimidating technology and packaging it as something genuinely fun and social. The Mad Libs approach solves real user experience problems while the recipe-sharing creates viral mechanics that could drive organic growth. Whether it can scale beyond the invite-only phase and compete with big tech's AI ambitions remains to be seen, but Things Inc. just proved there's still room for startups to reimagine how we interact with AI.
