An open-source personal health & fitness companion. Early stage, actively being built.
Repono comes from re- ("back") + pono ("I place, I put"). The verb reponere meant, across two thousand years of Latin:
- to put something back where it belongs — restore, replace
- to lay up, to store away — as in storing grain for the winter
- to restore strength or health — reponere vires, "to put the strength back"
It is the same root that gave English repository — a place where things are put back, kept, and found again.
Because the word carries the three things that actually matter in health, and they are the three things most fitness apps get wrong.
1. You do not get stronger while you train. You get stronger while you recover.
Training is the stress. Recovery is the adaptation. The muscle does not grow in the gym; it grows in the hours afterwards, when the body puts back what was taken. Repono is the correct name for the part of the process that most apps ignore — apps shout at you to do more, and stay quiet about rest.
2. Health is a stock, not a flow.
Every meal, every set, every night of sleep is a small deposit. The number that matters is not today's workout — it is the accumulated record. Reponere in Latin agriculture meant laying up the harvest so it would be there in winter. That is what a health record is: something put away now, so that a year from now you can look back and see what actually happened.
3. The goal is to put the body back, not to push it forward forever.
Most training language is about progress, PRs, pushing harder. But the reason anyone trains at all is to return to a state — to sleep well, to move without pain, to have energy at 4 p.m. The Latin word for that is not progredior ("I advance"). It is repono — I put it back.
We picked the name before we had a product. That is the normal order of things, but it means the name is currently a promise, not a proof. The word describes what we are trying to build: a quiet, cumulative, recovery-aware record of a person's health — not a gamified machine that punishes you for resting.
Repono is a personal health and fitness tracker. The current build is a WeChat mini-program backed by a small API. It records:
| Module | What it holds |
|---|---|
| People | The account and its baseline profile |
| Sessions | A training session: when, how long, what kind |
| Sets | The individual exercises within a session — load, reps, rest |
| Meals | What was eaten, roughly what it contained |
| Body metrics | Weight and other periodic measurements |
| Reports | Generated summaries that roll the above into a readable picture |
The design goal is the opposite of a leaderboard. The app should answer one question: "Is this person's body being restored, or is it being drained?" Everything else is decoration.
This is an early-stage project, still being optimized. Being specific about what that means:
Working
- Recording sessions, sets, meals and body metrics
- Generating summary reports
- Running as a WeChat mini-program against a live API
Not yet there
- The source is not published here yet. It currently lives in a managed cloud environment (Tencent CloudBase) and has not been extracted into a clean repository. This repo is the plan and the name, not the code.
- No public API documentation
- No automated tests
- No third-party security review
- Only a small number of internal test accounts
Why publish an empty-ish repo
Two reasons, both honest.
First, the name and the intent need a public home before the code is ready. A project with a name but no repository is a private idea. A project with a name and a repository is a commitment — it can be linked to, cited, and held to.
Second, the design decisions are the valuable part at this stage. The schema above, the recovery-first framing, and the deliberate refusal to gamify are more important right now than a working binary. When the code is published, it will be published against this document, so that any divergence between the promise and the product is visible rather than hidden.
| Stage | What it means |
|---|---|
| Now | Record and report. Get the data model right. |
| Next | Extract the source from the cloud environment into this repository. |
| Then | Publish the API surface and a self-host path. |
| Later | Recovery-aware recommendations — but only after the data is trustworthy. |
We will not ship recommendations before the underlying record is honest. A recommendation built on bad data is worse than no recommendation.
MIT. See LICENSE.
This project touches the most sensitive category of personal data there is. The current design position, stated plainly:
- The user's data belongs to the user. Export must always be possible.
- No advertising, no sale, no sharing with third parties. Not now, not later.
- No medical claims. Repono is a record-keeping tool. It does not diagnose, treat, or replace a clinician.
- No training on user data. The app's purpose is to serve the user, not to extract value from them.
These are not aspirational. If a future version cannot honour them, that version will not ship.
Trustbase Lab · 2026