Real Calorify simulator screenshots from November 2026 showing the paper-style Today dashboard with Nibbles and populated demo nutrient cards beside the Trends charts
← Project archive
Project 003 Software / AI

Calorify

I built a local-first iOS calorie tracker with AI coding agents, tried the features I wanted, and decided to archive the experiment.

Calorify’s simulator dashboard with the words Built with AI, Why I StoppedProject filmI Built a Calorie Tracker with AI… Then Decided to StopWatch on YouTube ↗

Why I started it

Calorify began as an experiment in building a calorie tracker with AI coding agents. I wanted to photograph a Canadian nutrition label and have the app fill in the values, rather than keep correcting inaccurate food entries in other apps.

What I built

I built the main screen around a daily calorie budget, with customizable cards for things like protein, sodium and fibre. Nibbles, the purple monster, eats the cookie as the day goes on and looks overstuffed when calories go well over budget.

Food search brought together saved foods, menu catalogues and Health Canada’s basic foods, with Open Food Facts for packaged products I hadn’t saved. I added barcode lookup, a label reader with a screen to check the values, and recipes made from other foods. The menu catalogues lived on my own home server and were downloaded by the app.

I also built iCloud household sharing for foods and recipes, plus files for sending a food, recipe or whole meal through AirDrop or Messages. Apple Health imports were there to bring entries from NetDiary into Calorify, alongside active-energy and weight settings. The trends screen used the same measurements as the dashboard.

Calorify Today dashboard with a calorie target and customizable nutrient cards
Today: a customizable view of the metrics that matter.
Calorify Trends screen with calorie intake and macronutrient charts
Trends: patterns over time, not just one logged day.

Keeping it local

One of the main design ideas from the start was keeping everything local. The diary stayed on the phone, label reading ran on the device, and there was no account or advertising.

The code was written with AI coding agents. My part was deciding what I wanted, testing the result, and correcting things as I went.

Why I’m archiving it

The label reader was the reason I started, but it never became reliable enough to save me much time. In real-world use it missed enough values that I sometimes ended up typing them in anyway.

Calorify was fun to experiment with, but building another calorie tracker isn’t something I feel interested enough in to keep going. I’m archiving the project here and keeping the video as a record of the ideas I tried.

Project timeline

Along the way

  1. Wrapping up the experiment

    I built Calorify with AI coding agents to try the features I wanted in a calorie tracker, starting with Canadian nutrition-label photos and keeping the diary and label reading on the phone. The label reader never became reliable enough to save me much time, and I decided not to keep building another calorie tracker. The wrap-up video records where the experiment got to before I archived it.

    Video update · Watch on YouTubeI Built a Calorie Tracker with AI… Then Decided to Stop
  2. The useful part turned out to be the view

    I started by trying to replace my calorie-counting app, but that space is already crowded. The more useful role is a custom view into data I already have. Community contributions still matter for foods that are harder to enter yourself — Dairy Queen, A&W, and other takeout with nutrition information online — while the app stays simple, local-first, HealthKit-connected, and able to back up custom foods to iCloud.

  3. Start: replace the calorie tracker I actually use

    I wanted a simpler way to add food — take a picture of a Canadian nutrition label, scan the barcode, or choose from a few basic foods I know. Other apps get those details wrong often enough that I was tired of correcting them. Keeping everything local, syncing with HealthKit, and backing up my custom foods to iCloud were part of the plan from the beginning.