MIT Students Create AI-Powered ‘Kitchen Cosmo’ Recipe Generator

MIT Students Develop AI-Powered Kitchen Device, Kitchen Cosmo

A team of students at the Massachusetts Institute of Technology (MIT) has developed Kitchen Cosmo, an artificial intelligence-powered device designed to generate recipes based on available ingredients and user preferences. The device, created by Jacob Payne and Ayah Mahmoud, aims to reimagine the role of AI in the kitchen, moving beyond simple automation to foster a collaborative cooking experience.

The project originated from an MIT course, “Interaction Intelligence,” which explores the design of a new category of AI-driven interactive objects known as large language objects (LLOs). According to Marcelo Coelho, associate professor of the practice in the MIT Department of Architecture and director of the Design Intelligence Lab, these objects are “physically situated and interact in real time with their physical environment.”

Payne and Mahmoud drew inspiration from the Honeywell Kitchen Computer of 1969, an early and ultimately unsuccessful attempt at bringing computing to the domestic kitchen. “It was an ambitious but impractical early attempt at a home kitchen computer,” Payne said. The students sought to create a functional prototype that addressed the limitations of its historical predecessor.

Kitchen Cosmo operates by scanning ingredients placed in front of its webcam. The AI then identifies these ingredients and generates a recipe tailored to user-defined parameters. These parameters include meal type, cooking time, difficulty level, number of portions, and dietary preferences. Users can also input their mood or desired “vibe,” with options ranging from nostalgic to celebratory.

The device’s interface is intentionally tactile, featuring dials, sliders, and switches rather than a touchscreen. This design choice, according to Payne, was intended to make the interaction with AI more visible and collaborative. “We wanted to challenge that default and ask what a different kind of interaction could look like,” he explained. “The goal wasn’t to make the interface seamless, but to make it tactile, playful and intentional.”

Once a recipe is generated, Kitchen Cosmo prints it on thermal paper via a built-in printer. The printed recipe can then be torn off and stored in a receptacle at the base of the appliance. The device was designed with a retro-futuristic aesthetic, featuring a bright-red body reminiscent of mid-century appliances.

The development process involved fine-tuning the AI to understand real-world cooking parameters, such as heating, timing, and temperature. The students also worked to ensure the AI could accurately recognize flavor profiles and spices from various cuisines. Testing involved taste-testing recipes generated by Kitchen Cosmo, with adjustments made based on the results.

While Kitchen Cosmo currently exists as a functioning prototype, both Payne and Mahmoud have ideas for future iterations. Payne envisions the device leveraging kitchen data to offer cooking tips, while Mahmoud is exploring ways to optimize the device for her thesis, potentially adding features such as multi-person instructions or a “learning mode” that provides guidance on using kitchen tools.

Mahmoud, who began her MIT education as a geologist, described the course as a pivotal moment in her design journey. “For the first time, in that class, I felt like I was finally drinking as much as I could and not feeling overwhelmed,” she said. “I see myself doing design long-term, which is something I didn’t think I would have said previously about technology.”

Kitchen Cosmo was first unveiled in August 2025, garnering attention in design publications such as Dezeen and Interesting Engineering.