Petitstats is a lightweight, privacy-friendly web analytics tool written in PHP. Most analytics tools need Node.js, Docker or a separate database server. Petitstats doesn't.
Petitstats was called Minilytics until October 2026. Existing installs update as usual from the dashboard and keep their data, snippets and settings: see Coming from Minilytics.
- Runs on a cheap Apache hosting plan, no VPS needed. It stores data in SQLite by default, and MySQL and MariaDB are also supported. Installation takes three steps. Just upload the archive and you are done.
- Bring your history with you. You can import your Umami and Matomo data from the dashboard or the command line. Importers for Google Analytics, Plausible and Simple Analytics are planned.
This is what the dashboard looks like. It is not because its lightweight that it can not be beautiful and functional
- PHP 8.1 or later with the
sqlite3andpdo_sqliteextensions enabled - Apache or LiteSpeed work out of the box; for Nginx or Caddy, apply the web server configuration
- For MySQL or MariaDB: the
pdo_mysqlPHP extension and a database/user withCREATE,ALTER,INDEX,SELECT,INSERT,UPDATEandDELETEpermissions - A writable data directory outside the web root (default:
petitstats-data/next to the web root, or setPETITSTATS_DATA_DIR) - Outbound HTTPS access and
zlibsupport (to download the local MaxMind GeoLite2 City geolocation database)
Note: The zip extension is only needed if you import historical data from external services.
Petitstats is easy to install. Download the latest production archive from GitHub Releases and extract it into your website root directory:
curl -LO https://github.com/axthauvin/petitstats/releases/latest/download/petitstats.tar.gz
tar -xzf petitstats.tar.gz && rm petitstats.tar.gzThats it!
Once its installed, navigate to https://your-domain.com/dashboard/ to create the initial administrator account.
SQLite is the default (because it's lightweight and doesn't require a separated server), but you can change it any time !
To use a managed database, open Settings → Database, select MySQL or MariaDB, enter the host, port, database name, username and password, then use Test connection before saving. The test can create the named database when it is missing if the database user has the CREATE permission. Existing SQLite analytics are not copied automatically.
Data is stored outside the web root so redeploying or re-extracting the archive can never overwrite or expose it.
Clone the repository, install the PHP dependencies with Composer and start the development server:
cd app
composer install
composer serve # or php -S localhost:8080 -t app scripts/dev-router.phpOpen http://localhost:8080/dashboard/ for the dashboard and http://localhost:8080/ for the landing page. The development router (scripts/dev-router.php) overlays landing/ on top of app/. Without Composer, run php -S localhost:8080 -t app scripts/dev-router.php from the repository root.
Release archives already bundle vendor/, so Composer is only needed when running from a clone of the repository.
Want to contribute? Read the contributing guide for the coding standard, checks and commit convention.
- Tracking and custom events
- Privacy and data protection
- Operations, backups and storage
- Importing data from other services
- AI assistants (MCP)
Node.js is used only to run unit and smoke tests during development. Node.js is not required to run Petitstats in production.
node --test app/tests/*.test.jsThis project is licensed under the MIT License - see the LICENSE file for details.