#foss-roundup#github#small#niche

Under-100-stars FOSS nutrition finds

Small projects on GitHub that won't show up in any roundup but are doing something interesting.

By Alex Brennan · Published March 29, 2026

Why under 100 stars

The big FOSS trackers (OpenNutriTracker, Waistline, OpenScale) get the attention. Below them is a long tail of single-purpose tools that solve narrow problems well. We use several of these in our personal stack and none of them have the visibility to make annual lists.

Five we’ve actually used in 2025–2026:

1. nutrient-cli

A bash script + Python helper that wraps the USDA FDC API for one-off lookups from a terminal. About 200 lines total. Useful when:

You’re cooking and want to fact-check a calorie claim from a recipe.
You want to add nutrition to a personal shell prompt or Tmux statusline.
You’re scripting something against the FDC API and want a working reference.

Star count: ~40. Last commit: 4 months ago. Single maintainer who is actively responsive on issues.

We aliased it to cal in ~/.zshrc and use it daily.

2. recipe-cost

A self-hosted Flask app that takes a recipe (yield + ingredient list with quantities), looks up macros from USDA, and outputs:

Per-serving macros
Per-serving ingredient cost (using user-supplied unit prices)
Cost per 100 kcal, cost per gram of protein

The protein-per-dollar metric is what we use it for. We’ve spreadsheet-tracked grocery prices for years and the script makes the analysis trivial.

Star count: ~70. Last commit: 6 weeks ago. Two contributors.

3. mealtemplate-export

A small Python tool that takes a markdown file with a structured recipe block and emits OpenNutriTracker-compatible custom-recipe JSON. Solves the “I want to define recipes in plain text and import to my tracker” use case.

Star count: ~15. Last commit: 8 months ago. The maintainer doesn’t update it often but the tool is finished — it does one thing.

We use this to keep our recipe definitions in a private git repo (markdown, version-controlled) and pull them into ONT when needed.

4. fdds-survey-extract

A Postgres-loader specifically for FNDDS (Food and Nutrient Database for Dietary Studies) data, separate from the larger FDC bulk loader. FNDDS has slightly different shape than the rest of FDC and most generic loaders skip it.

Star count: ~25. Last commit: 3 months ago. One maintainer, an academic.

We use this for a personal “cooked-as-prepared” lookup that the standard FDC loader doesn’t surface as cleanly.

5. weight-trend-tools

A Python notebook + small CLI for analysing exported weight data. Takes a CSV of (date, weight_kg) and produces:

LOESS-smoothed trend
Weekly delta
Confidence interval on the rate of change
Outlier detection for measurement glitches

Star count: ~80. Last commit: 5 weeks ago. Three contributors.

We’ve used this for two years to clean up our OpenScale exports before importing into a tracker. The outlier detection alone has saved us from at least one panicky over-reaction to a single bad scale read.

Honourable mentions

A handful we’ve used briefly but don’t currently:

A FastAPI wrapper around the FDC API that adds a local Redis cache. Useful if you’re hitting FDC rate limits in dev. ~50 stars.
A Markdown-to-iCal “meal calendar” generator that puts your planned meals on your calendar app. ~30 stars. Niche, but cute.
An import-from-MFP-export tool that takes MFP’s GDPR-export JSON and converts it to the schema OpenNutriTracker wants. ~60 stars. Useful if you’re switching.

Why these matter

The big trackers have to be general-purpose. The small projects can be opinionated and sharp. Most of them are 200–500 lines of code; if the maintainer disappears, you can fork and read the whole thing in an afternoon.

The other reason to know about them: they often die. Of the ~20 sub-100-star nutrition projects we evaluated for this piece, the five above are still active. The other 15 are stale or abandoned. That’s the survivorship rate.

How to find more

A few search strategies that work:

site:github.com calorie tracker stars:<100
site:github.com nutrition stars:<100 language:python
F-Droid’s “new since last release” feed
The OFF and FDC API guides, which sometimes link to community tools

If you find one we missed, email us and we’ll add it to the next roundup.

What not to do with small projects

Don’t put your only data in them. Always also export to a tracker that can re-import.
Don’t expose them to the public internet without reading the code first.
Don’t expect 24/7 maintenance. Most are someone’s evening project.