mljar-supervised
Automated machine learning Python package for tabular data with feature engineering and automatic documentation.
Build, tune, and explain tabular ML pipelines with automated documentation and web apps.
mljar-supervised is an automated machine learning Python package designed to save data scientists time on tabular data tasks. It handles data preprocessing, algorithm selection, hyperparameter tuning, and generates automated Markdown reports.
What it is
The package trains multiple models including linear models, random forests, XGBoost, LightGBM, CatBoost, and neural networks, combining them via ensembling and stacking. It computes baseline models, handles missing values, creates golden features, and provides SHAP-based model explanations.
Who it's for
It suits data scientists and analysts looking to automate repetitive pipeline construction and hyperparameter tuning for tabular data. It does not suit unstructured data tasks like raw image or audio processing.
Key features
Built-in Modes
Choose between Explain, Perform, Compete, and Optuna modes depending on your project goals.
Automatic Documentation
Generates detailed Markdown reports for each ML model and analysis step.
Web App Generation
Automatically generate prediction web apps using Mercury to share with domain experts.
Fairness Aware Training
Includes options for fairness-aware model training on tabular data.
Up and running in 3 steps
Our honest verdict
What's great
- Generates interactive prediction web apps directly from trained models.
- Provides detailed Markdown reports with visual explanations and SHAP values.
- Supports multiple built-in operating modes ranging from explainability to competition-level stacking.
- Handles missing values and advanced feature engineering automatically.
Where it falls short
- Computationally intensive when running Compete or Optuna modes across many algorithms.
- Limited strictly to tabular data formats.
- Requires Python environment setup and dependency management.
Pricing & license
Free under its license, with the only cost being your own compute resources. MLJAR also offers a commercial hosted option called MLJAR Studio.
Open remote AI-training roles
874 open remote AI-training roles, updated November 15, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
What data formats are supported?
It is designed specifically to work with tabular data.
Can I generate web apps for non-technical users?
Yes, you can automatically generate and publish web apps using the integrated Mercury tool.
