A small independent app studio out of Finland released a new nutrition-tracking app on September 17, 2026, that tries to answer a question calorie counters have long struggled with: not just how much you ate, but what kind of food it actually was. Pureo, built by the studio Sober Bar and led by executive director Robert Winter, launched simultaneously on iPhone and Android with a pitch aimed at people who feel calorie-only tracking has plateaued as a behavior-change tool.
Beyond the Calorie Count
“Calorie counting was a revolution in nutrition tracking, but it only tells part of the story,” Winter said in the app’s launch materials. His argument echoes a growing body of nutrition science suggesting that two meals with identical calorie counts can have wildly different effects on the body depending on sugar content, fiber, and how heavily processed the ingredients are. Pureo’s core feature lets users photograph a plate of food or type a short description, and its AI model estimates calories, protein and other macronutrients — a photo-recognition approach now standard across several nutrition apps — while adding two metrics competitors typically skip: an ultra-processed-food score and a dedicated sugar tracker.
How the AI Analysis Works
When a user snaps a photo of a meal, Pureo’s model attempts to identify individual food items, estimate portion sizes and cross-reference likely ingredients against processing categories, flagging items that fall into what nutrition researchers call the NOVA classification system’s highest-processing tier — think packaged snacks, sugary drinks and reconstituted meat products, as opposed to whole or minimally processed foods. The app then rolls this into a daily tracking score meant to give users a single number reflecting both quantity and quality of what they ate that day, alongside more granular breakdowns they can drill into.
Riding a Broader Shift in Nutrition Science
Pureo’s launch lands amid several years of mounting research and public attention on ultra-processed foods, following studies published by researchers including NIH’s Kevin Hall linking heavily processed diets to overeating and metabolic problems even when calorie counts are matched against minimally processed alternatives. Public-health messaging in the U.S., U.K. and elsewhere has increasingly shifted from a pure calories-in-calories-out framework toward food-quality guidance, and Pureo is betting that consumer apps have lagged behind that shift. Existing giants in the category, such as MyFitnessPal and Cronometer, have added AI photo-recognition features over the past two years but have been slower to build dedicated processing-level scoring into their core experience.
A Crowded, Skeptical Market
Nutrition and technology reviewers have grown increasingly cautious about AI-based food-photo analysis after several high-profile apps were shown to misidentify portion sizes or ingredients hidden under sauces and garnishes, sometimes undercounting calories by wide margins. A nutrition expert who tested several AI diet-tracking apps in September 2026 found meaningful accuracy gaps between products, underscoring that visual estimation of food remains a genuinely hard computer-vision problem — a plate of pasta with cream sauce looks very different in calorie terms than one with tomato sauce, but both can appear nearly identical to a camera. Pureo has not yet published independent accuracy benchmarks for its own model, and it remains untested by third-party nutrition researchers.
Business Model and Availability
Sober Bar has not disclosed pricing details for Pureo, though the company describes it as free to download on both the App Store and Google Play, with the company’s site, pureo.app, serving as the main hub for feature updates. As an independent studio rather than a venture-backed startup, Sober Bar faces the same customer-acquisition challenge every new nutrition app confronts: convincing users to switch from tools they may have used for years, or to add yet another app to a phone already crowded with health trackers.
Why It Matters and What’s Next
The bet behind Pureo reflects a broader pattern in consumer AI: companies are trying to differentiate not by building a faster or cheaper version of an existing tool, but by using AI to surface a dimension of a familiar problem that older, manual-entry apps couldn’t easily capture — in this case, food quality rather than just quantity. Whether that differentiation sticks will depend on whether Pureo’s AI model proves accurate enough, over months of real-world use, to earn user trust on judgments as specific as “ultra-processed” versus “minimally processed.” If it does, expect larger nutrition-app incumbents to fold similar scoring into their own products within the next product cycle rather than cede the category to a small independent studio.
Photo: Amar Preciado / PEXELS via Pexels
