COACHINT in short

What it is, who it is for, and how it works.

In one sentence: COACHINT reads an athlete's own training data, learns what is normal for them, and explains every day what changed and why it matters. A coach sees the same explanations for the athletes who chose to share with them.

The problem

An athlete today uses a watch, one or two apps, maybe a coach's spreadsheet. Each shows numbers. None says, in plain words, which of those numbers matter this week, for this person.

A coach with twenty athletes has the same problem twenty times over: who needs attention today, and why?

What COACHINT does

1. Brings the data together. Garmin, Polar and Strava exports, watch files (FIT, TCX, GPX) and Intervals.icu land in one history. A workout that arrives twice is counted once.
2. Learns your normal. After about five weeks it knows your usual training load, resting heart rate, heart-rate variability and sleep.
3. Says what changed. Each week it lists what is unusual compared with your own last weeks: load climbing too fast, missed sessions, poor recovery, a better pace for the same effort. Each signal shows the numbers behind it.
4. Reads your notes. A note like "knee sore since km 8" or "rough week at work" becomes a signal too.
5. Answers questions. Ask about your week in any language. Every number in the answer is checked against your data.
6. Tells you every morning. A short digest by Telegram, email or in the app, with Useful and Dismiss buttons so it learns what helps.

Who it is for

• Athletes who want to understand their own training without becoming data analysts.
• Coaches who want to see, for the athletes who agreed, who needs attention first and why. The coach view reuses the same engine as the athlete's.

What it believes

• You against yourself. No generic norms: every alert compares you with your own history.
• Numbers come from the data. The AI explains; it does not invent figures.
• Your data is yours. Nobody sees it unless you agree. You can download all of it, or delete your account, at any time.
• Honest sources. Data arrives only through routes the providers allow for this use: your own exports, Polar with your consent, Intervals.icu with your own key. No Strava API.
• Not a doctor. Signals are prompts to look closer, not diagnoses.

Where it stands

COACHINT is an early, private project: a proof of concept tried in a small pilot by a group of friends and one or two coaches, with their own training data. It is not a product you can sign up for, and whether it becomes one or an open-source project is not decided yet. The feature summary lists what is built and what has not run with real data yet. The simple manual shows how to use it; the tutorial explains every screen.