About
Built by someone who
couldn't find the right app.
Henry Wagner
Builder · University of Michigan
Over the past year, I've become deeply invested in the gym and my health. Like a lot of people, I turned to macro tracking to understand what I was eating — and like a lot of people, I got frustrated fast.
Every app I tried fell into one of two buckets: super manual entry (search, scroll, tap, repeat) or overly AI-based apps that make guessesabout nutrition data. Neither felt right. Manual logging killed my flow while cooking. And I didn't trust apps that were just guessing my macros.
I wanted something that worked with me while I cooked. Something I could talk to naturally, that would pull from verified data sources, and that supported multiple ways to log — voice, barcode, camera, or just typing it in. I wanted an app that met me where I was, not one that forced me into a single workflow.
That's how Kitchen Mode was born — a hands-free, voice-first logging experience powered by AI that parses what you say but never fabricates nutrition data. You cook, you talk, and your macros are logged.
Then I realized a Bluetooth scale could change everything. If Kitchen Mode already knows what you're eating, and a scale knows how much, you get incredibly accurate logging with almost zero effort. No more eyeballing portions. No more guessing if that was one tablespoon or two.
The longer-term vision goes further. I'm focused on building a reputable database of foods with barcode data and complete nutrition information. The goal is to make collecting great data really easy — and then see how that data can help people. Anonymized, high-quality nutrition data has the potential to power assistive health models that help people understand patterns in how they eat and feel.
But that starts with getting the basics right: making it dead simple to log what you eat, with data you can actually trust.
University of Michigan
Dialed is being developed at the University of Michigan, combining research in human-computer interaction, mobile computing, and health informatics. The project draws on faculty guidance in AR/HCI and a rigorous engineering approach to building consumer health tools that are both technically sound and genuinely useful.
Built with
SwiftUI
Native iOS
Gemini AI
Voice parsing
USDA FoodData
Nutrition data
CoreBluetooth
Scale integration
Fastify
API server
PostgreSQL
Database
WebSocket
Real-time streaming
Prisma
Data layer