Self-hosted service
wger-project/wger avatar
wger-project/wger

wger keeps workouts, nutrition and photos on a server you run

Self hosted FLOSS fitness/workout, nutrition and weight tracker

7,268 stars1,070 forksPythonAGPL-3.0

At a glance

What is it?
wger is a self-hosted workout and nutrition manager: Django and htmx on the server, a community exercise database, Celery and Redis behind it, and a REST API that mobile clients call.
Who is it for?
Adopt wger if you want training logs, body weight, diet entries and progress photos to live on infrastructure you control, and if you are willing to run Redis, Celery workers and a Django upgrade cycle instead of tapping a settings screen. Do not adopt it for wearable or heart rate data, since nothing in the project lists that, or for a food database you can rely on offline, since nutrition data comes from Open Food Facts.
Can I use it commercially?
Yes, with strict conditions. AGPL-3.0 is a network copyleft licence: if people use a modified version over a network, for example as a hosted service, you must offer them its source code under the same licence.
Is it still maintained?
Yes. The repository last received commits 2 days ago.
What is it written in?
Mainly Python, according to GitHub's language statistics.

Answers come from the project's GitHub data, last synced on November 13, 2026, and from our analysis. They are not legal advice.

From daily snapshots

Star history

  • wger-project/wger7,268 ★ · +360
Star history
Repository2026-09-132026-11-13
wger-project/wger6.9k7.3k

360 Stars in the last 61 days, measured from the daily GitHub snapshot. Compare with other projects

Editorial analysis

A Django app that outgrew the workout log

wger (ˈvɛɡɐ) is a free workout and fitness manager, and the current feature set reaches further than that name. Alongside routines and training logs it tracks diet plans, body weight and custom measurements, keeps a gallery of progress photos, and runs a shared exercise and ingredient database. The premise is that all of it sits on an instance you control instead of on somebody else's account.

The exercise wiki is the part that separates it from a private log. Exercises are built in and users can contribute to them, and the project aims for a comprehensive list of exercises and ingredients. Nutrition entries draw their food information from Open Food Facts, so coverage and gaps both come from that database.

Multi-user support exists and is described as basic gym management, which is enough for a club sharing one instance and not much more. A live system runs at wger.de if you want to look at the interface before hosting anything.

Progression rules carry the training logic

Routines are the unit of work, and one feature inside them is called out specifically: automatic weight progression rules. That is the piece a plain logbook leaves to memory, and it is the reason wger is more than a table of sets and reps.

The README does not describe how a rule is written; the documentation at wger.readthedocs.io is the place to look.

The dependency list shows where else the app invests. icalendar is a required package. django-simple-history and django-activity-stream are both present, so edits to entries leave a history trail and produce an activity feed. On the reporting side, the tracked quantities named in the feature list are diet plans, body weight and custom measurements, and custom measurements are the hook for anything the project does not model by default.

Docker Compose is the install, and the file lives next door

The documented install is one command.

bash
docker compose up -d

The compose file itself is not in this repository. The overview points to a separate docker compose file under the wger-project organisation and to the installation page at wger.readthedocs.io/en/latest/installation/docker.html for the detailed steps, so the two pieces needed on day one live in different places.

The Python side is heavier than that single command suggests. The floor is set in pyproject.toml:

toml
requires-python = ">=3.12"

Builds go through hatchling, the project is classified as Django based, and Django is required at ~=6.0.7 with the argon2 extra. Celery is required with the redis extra and django-redis is a separate dependency, so background jobs and caching assume a Redis service exists. uv.lock pins the Python side, while manage.py and a settings directory sit at the repository root in the conventional Django arrangement.

Celery, Prometheus and object storage behind the pages

Read the dependency list and the shape of a deployment becomes clearer. celery[redis] and flower are both required, so asynchronous work and the tool that monitors it arrive as a pair rather than as something you add later. django-prometheus is required as well, which puts metrics inside the default stack and makes scraping part of operating the app rather than an optional extra.

The photo features carry their own requirements. easy-thumbnails handles image resizing, and django-storages with boto3 covers object storage, so a progress gallery can live in S3-compatible storage instead of on the container filesystem. That is a deployment choice the overview leaves open. django-environ reads configuration from the environment rather than from a checked-in file.

Translations run through Weblate and the interface is available in many languages, translated by the community. There is no architecture document in the tree, so the dependency list and the settings directory remain the honest places to look before sizing a host.

OpenID Connect, MFA and locked accounts in the login stack

Authentication is where a self-hosted instance carries most of its risk, and here the dependency list is unusually explicit. django-allauth is required with the mfa and idp-oidc extras, so multi-factor authentication and OpenID Connect single sign-on are built in rather than bolted on afterwards. The API side uses djangorestframework with simplejwt and its crypto extra, so tokens are signed rather than opaque.

Brute force handling is not optional either. django-axes is required with ipware, and django-recaptcha is a separate dependency, which means failed logins are counted per address and public forms are gated. Password hashing comes from the argon2 extra on Django, and django-cors-headers is present for browsers calling the API from another origin.

The tree also carries SECURITY.md, so disclosure policy has a home, and the overview says nothing about audits. Nor does it promise a deployment hardening guide. For an instance open to the internet that is worth noting: the app brings rate limiting and lockouts, and leaves firewall, TLS and backup decisions to the operator.

htmx and Bootstrap, compiled from SCSS

The front end is deliberately small. package.json requires htmx.org 4.0.0, bootstrap 5.3.8, datatables.net-bs5 and @popperjs/core, which reads as server-rendered Django templates with a table plugin rather than a single-page application. Styling starts as SCSS and is compiled into a stylesheet the repository keeps:

bash
npm run build:css:sass

That script runs sass over wger/core/static/scss/main.scss and writes wger/core/static/bootstrap-compiled.css. One dev dependency, @wger-project/react-components, is pinned to the exact version 26.10.4, so shared components arrive as a numbered release rather than being tracked from source, and packageManager is set to npm 10.5.0 to match.

A mismatch sits in the same file. package.json declares version 2.6-alpha1 while release 2.7 was published on 2026-09-03, so the JavaScript asset manifest does not track the Python release numbering. Harmless for a user, relevant to anyone pinning a deployment against a tag.

Three licences on one repository

Licensing is split three ways, and the split matters if you build on top. Application code is AGPL-3.0-or-later, declared in pyproject.toml with LICENSE.txt and AUTHORS.md listed as licence files. Exercise and ingredient data is Creative Commons, with the overview pointing at individual entries rather than promising one blanket grant, so an export carries per-entry terms. Documentation is CC-BY-SA-4.0.

For a gym or club hosting an instance for members, the code licence is the part to read first, because AGPL obligations reach a modified version that others interact with over a network. For a commercial product that wants the exercise data and little else, the data terms matter more than the code terms. Neither point is legal advice, and the per-entry Creative Commons note is the sentence to read before redistributing any data export.

Release cadence is regular: 2.5 on 2026-04-15, 2.6 on 2026-06-17 and 2.7 on 2026-09-03, with the last push dated 2026-10-29. The default branch is master, which trips up anyone who clones by convention and expects main.

What the feature list leaves out

Several things you might expect appear nowhere in the overview or the dependencies. The README mentions no wearable or heart rate integration and no barcode scanning or camera-based food logging. The exercise and ingredient data is community-contributed, so coverage depends on what other users have added, and it does not sit under one licence.

The mobile apps are separate projects. Android, F-Droid and Flathub builds are listed under the package id de.wger.flutter, with an iOS app on the App Store, so a phone talks to your instance through the REST API rather than carrying a local copy of the database. That moves the API surface, JWT authentication and the availability of your own server onto the phone side of the equation.

Two limits are aimed at operators rather than athletes. The compose file lives in another repository, so a self-hosted deployment cannot be pinned to a tag in this tree, and Python 3.12 is the floor, which rules out any host pinned to an older interpreter.

Editorial conclusion

Adopt wger if you want training logs, body weight, diet entries and progress photos to live on infrastructure you control, and if you are willing to run Redis, Celery workers and a Django upgrade cycle instead of tapping a settings screen. Do not adopt it for wearable or heart rate data, since nothing in the project lists that, or for a food database you can rely on offline, since nutrition data comes from Open Food Facts. Verify three things before the first workout: which commit of the separate docker repository the compose file should come from, whether the exercise and ingredient entries you plan to redistribute carry the Creative Commons terms the README points at, and whether the REST API plus JWT authentication is enough for the mobile clients you intend to hand out. The last push was 2026-10-29 and release 2.7 shipped on 2026-09-03.

Frequently asked questions

What is wger?

wger is a free, self-hosted workout and fitness manager written in Django. It tracks routines with automatic weight progression rules, diet plans, body weight, custom measurements and progress photos, and it ships a community exercise and ingredient database that users can contribute to.

How much does wger cost?

The application code is AGPL-3.0-or-later, so the software itself is free to use and self-host. The exercise and ingredient data is Creative Commons with terms set per entry, and the documentation is CC-BY-SA-4.0.

What are the benefits of wger?

The project lists self-hosting with Docker, a REST API for third-party integrations or automations, multi-user support with basic gym management, community translations through Weblate, and mobile clients on Android, iOS, F-Droid and Flathub. Running your own instance keeps the training log, photos and measurements on infrastructure you control.

Official sources

  1. License: AGPL-3.0
  2. Project website
  3. README
  4. Releases
  5. wger-project/wger on GitHub
Add this badge to your README

If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.

Add this badge to your README

markdown
[![Hysen Labs](https://hysenlabs.com/badge/wger-project-wger.svg)](https://hysenlabs.com/projects/wger-project-wger)

Star history chart for your README

markdown
[![Star history](https://hysenlabs.com/chart/wger-project-wger.svg)](https://hysenlabs.com/projects/wger-project-wger)

Was this useful?