GymTracker

GymTracker is a private, mobile-first app for tracking workouts, plans, body measurements, goals and supplements. It is a monorepo with a Fastify API, a React client and a PostgreSQL database, meant to run with Docker Compose on a home server and be reached over Tailscale.

View code on GitHub Run it locally

01 · Problem

What needed solving

Workout apps either keep your data in someone else's cloud or don't track everything in one place. I wanted plans, sets, measurements, goals and supplements together, fast to use from a phone in the gym.

02 · Solution

How it works

npm workspaces with apps/api (Fastify 5, Zod, Prisma), apps/web (React 19, TanStack Query, Tailwind) and packages/shared. The PostgreSQL 17 schema is managed with Prisma migrations; Docker Compose runs the database, the API and an Nginx-served frontend, and applies migrations on start.

03 · Key features

What it does

Workouts with sets, weights, reps and notes; training plans by weekday
Calendar of planned, completed and skipped workouts
Body measurements (weight, body fat, circumferences, steps) with trend charts
Goals, a daily supplement checklist and export of any date range to JSON, CSV or TXT
Mobile-first UI: bottom navigation on phones, side panel from 900 px

04 · Tech stack

Built with

TypeScript
Fastify
Prisma
PostgreSQL
Zod
React
Docker

Status: Ongoing
Code: github.com/KacperBlok1/analiza-silownia
Demo: Not hosted yet · run locally

05 · Challenges

What was tricky

Consistent API errors
One Fastify error handler maps Zod validation errors to 400, a missing Prisma record (P2025) to 404 and a unique-constraint clash (P2002) to 409, so the route handlers don't repeat the same try/catch.
Schema migrations
The data model grew through four versioned Prisma migrations (initial schema, workouts, progress features, calendar history), applied automatically when the container starts.
No login
There is no login on purpose: the app runs on a LAN and over Tailscale, and PostgreSQL is only reachable inside the Docker network. Exposing it to the internet would need an auth layer first.

06 · Screenshots

How it looks

GymTracker desktop view with today's planned workout — Today: the planned workout and a start buttonOpen full-size image ↗

Active workout screen with the exercise, sets and a rest timer — Active workout: current exercise, sets and rest timerOpen full-size image ↗

Monthly calendar with planned workouts and their status — Calendar with the plan for the coming weeksOpen full-size image ↗

Progress page with weight, body fat and steps charts — Progress: latest measurements and chartsOpen full-size image ↗

Try it

Run it locally

No hosted demo yet. The source and README are on GitHub, and these commands start it locally from the repo root.

GitHub repository

terminal

$ cp .env.example .env
# set POSTGRES_PASSWORD
$ docker compose up -d --build
$ docker compose exec api npm run prisma:seed -w @gymtracker/api
# web: http://localhost:8088