Why Cronometer wasn't the answer
I tried the more conscientious commercial app and ended up self-hosting anyway.
I want to be fair to Cronometer.
After I left MyFitnessPal, the obvious closed-source replacement was Cronometer. Better privacy posture, better micronutrient depth, well-regarded among the kinds of people who actually take nutrition seriously. The Cronometer subreddit is, at the time of writing, one of the more thoughtful nutrition-app communities online — there’s none of the diet-evangelism, none of the vitamin-supplement hyping, just people with spreadsheets discussing whether they’re getting enough vitamin K2.
I subscribed. I used Cronometer Gold for four months in late 2024. I cancelled.
This piece is about why.
What Cronometer got right
The core product is excellent. Better than MFP across the board on the things that matter to me:
- Database depth on micronutrients. They show 80+ nutrients per food where MFP shows 12. The data is curated rather than crowdsourced — a smaller catalogue but cleaner.
- Privacy posture. No behavioural-ad tracking in observed traffic. Their privacy policy explicitly disclaims selling data. (See the policy walkthrough for the long version.)
- The app itself feels well-built. Charts load fast. The diary view is opinionated in the right way. The micronutrient warnings on under-consumption are actually useful.
- Cancellation works without confirmshaming. I clicked “cancel,” I confirmed, that was it.
Cronometer’s team has, by every signal I can read, been making the right trade-offs for a long time. The app has gotten better, not worse, over the years. I have nothing against it.
Why I cancelled anyway
Three reasons, all about the structure rather than the product.
1. The renewal calendar
Subscription software, even good subscription software, puts you on a calendar that isn’t yours. Once a year, an automatic charge. The day you forget to check is the day the charge happens. The choice to keep using becomes the default; the choice to leave becomes the active decision.
This is a small thing. It’s also a thing I dislike enough about every subscription I’m on that I’ve systematically reduced them. Cronometer was on the list.
2. The migration risk
If you build a workflow on Cronometer Gold, you’re building on Cronometer’s continued existence as a company. They’re independent now. They could be acquired. The acquirer might preserve the practices; or might not. The contract you signed with the 2024-Cronometer is not necessarily the same as the contract you’ll be inside in 2030.
This is what the Francisco Partners / MyFitnessPal experience taught me. Good ownership today doesn’t guarantee good ownership tomorrow. The only fully-stable answer is software you control.
3. The micronutrient mirage
Cronometer’s micronutrient depth is its differentiator. I subscribed mostly because of that. Four months later I’d checked my zinc and my magnesium and my vitamin K2 and I’d never gone back. I had calibrated the diet I’d already built. I didn’t need the daily granularity Cronometer offered for the diet to keep working.
This is a personal observation, not a criticism. For someone who does need to dial in micronutrients constantly — a cancer survivor managing vitamin K interactions, an athlete on a strict cutting cycle — Cronometer’s depth is genuinely useful. For me, after the initial calibration, it was over-engineered.
What replaced it
The same FOSS stack I’d already been running parallel to Cronometer:
- OpenNutriTracker on the phone
- Open Food Facts for barcode lookups
- A self-hosted USDA FDC mirror in Postgres for the occasional micronutrient check
- A Python script for the targeted use cases I’d built around (workouts, recipes)
The micronutrient surface in the FOSS stack is shallower than Cronometer’s. I check it monthly via a Jupyter notebook against the Postgres mirror, not daily through an app screen. The reduction in attention is, on net, fine.
What I’d say to someone considering Cronometer
If you’re choosing between MyFitnessPal and Cronometer: Cronometer. It is materially better on every axis I care about, including privacy.
If you’re choosing between Cronometer and self-hosting: it depends on whether you have, or want, a homelab. The cost-benefit favours Cronometer ($59/yr vs an evening of setup + the long tail of running it) for someone whose time is worth more than the savings. The cost-benefit favours self-hosting for someone who already runs servers for other things.
I run servers for other things. I went FOSS.
What I’d say to the Cronometer team
They probably won’t read this. If they do: you’ve built one of the better commercial calorie trackers. Your privacy posture is a credit to the field. The reason I left has nothing to do with what you’ve done; it has to do with the structural fact of being a subscription customer of a company I don’t control.
If you ever shipped a self-hostable tier — even a paid-once, run-on-your-own-Postgres model — I’d subscribe.
Until then I’ll keep my homelab and my OFF mirror, and I’ll quietly recommend Cronometer to people for whom self-hosting is not the right answer.
Footnote, August 2025
A reader emailed asking whether the FOSS stack ever feels like deprivation. Yes, sometimes. The first month felt rough. By month three it felt routine. By month six it felt clearly correct. The deprivation was real, then transient, then absent. Most habit-changes work like this; the literature on it is a Wikipedia rabbit hole worth the time.