Flowtrack

A private, self-hosted period tracker. It logs period days with flow and symptoms, predicts the next six periods, and estimates the fertile window. It installs on a phone home screen as a web app.

The app has no runtime dependencies beyond the Python standard library, and it stores everything in one SQLite file. Nginx and Authelia provide the access boundary.

Predictions

• Periods are runs of logged period days. Spotting doesn't count, and one missing day inside a run doesn't split it.
• The cycle length is a weighted average of up to the last 12 cycles, with recent cycles counting more. Cycles shorter than 15 days or longer than 60 are left out because they usually mean a log was missed.
• Each prediction is a range of ± the cycle's standard deviation (at least 2 days, or 4 with fewer than three cycles). The range widens with √n for later cycles.
• Ovulation is estimated 14 days before the next period. The fertile window runs from 5 days before ovulation to 1 day after. These are estimates, not contraception.
• Once the expected day passes, the next period is expected from tomorrow, and later predictions move with it.

The logic lives in static/cycle.js and is tested in tests/test_cycle.js.

Access

Nginx must overwrite Remote-User with the Authelia username. Users listed with --editor can log and export. Users listed with --viewer see period days and predictions only. They also see flow and symptoms if the owner turns on sharing in Settings. Everyone else gets 403.

Run locally

python3 app.py --database /tmp/flowtrack.sqlite3 --editor me --dev-user me

Then open http://127.0.0.1:8093.

NixOS module

services.flowtrack = {
  enable = true;
  editors = [ "kayla" ];
  viewers = [ "fred" ];
  backup.directory = "/mnt/disk2/backups/flowtrack";  # nightly, keeps 30
};

python3 app.py --database DB --backup-to DIR --keep N takes a consistent snapshot by hand.