kedro

What it is and what it does

Kedro is a framework for structuring data engineering and data science projects using software engineering best practices. It provides a project template based on Cookiecutter Data Science, a data catalog for managing data connectors across multiple file formats and storage systems, automatic pipeline dependency resolution, and visualization via Kedro-Viz. The framework emphasizes reproducibility, modularity, and team collaboration by moving away from Jupyter notebooks and ad-hoc scripts toward maintainable, reusable code.

The package includes 17 runtime dependencies covering configuration management (OmegaConf, Dynaconf), CLI tooling (Click), templating (Cookiecutter), file system abstraction (fsspec), version control integration (GitPython), and output formatting (Rich). It supports Python 3.10 through 3.14 and is actively maintained by the Kedro product team and open-source contributors.

Use it for:

• Build reproducible machine learning pipelines with automatic task dependency tracking and data versioning for file-based systems
• Structure team data science projects with standardized layouts, coding standards, and test-driven development practices
• Deploy data workflows to distributed systems including Argo, Prefect, Kubeflow, AWS Batch, and Databricks
• Manage complex data transformations across multiple file formats and cloud storage backends via a unified data catalog
• Visualize and debug pipeline execution flow and data lineage using Kedro-Viz integration

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Kedro is a Python framework for building production-ready data engineering and data science pipelines with automatic dependency resolution, data catalog management, and built-in support for reproducibility and modularity.

Yes. Kedro is production-stable (Development Status 5), actively maintained, has no known vulnerabilities, low install friction, and a permissive Apache 2.0 license. It is well-suited for teams building reproducible, modular data pipelines at scale. Install it if you need structure and best practices for data engineering or data science projects beyond notebooks.

Install

kedro on PyPI

pip install kedro

uv add kedro

poetry add kedro

Installing kedro

Before you install

Low install friction with a pure-wheel distribution and 17 well-established runtime dependencies. Active maintenance with a recent release 46 days ago and ongoing commits; the project has 10953 stars and is hosted by the LF AI & Data Foundation.

License in practice

Apache 2.0 permissive license allows commercial use, modification, and distribution with minimal restrictions, making it suitable for proprietary and open-source projects alike.

Quickstart

pip install kedro

from kedro.pipeline import Pipeline, node
from kedro.io import DataCatalog

# Define a simple pipeline node
def process_data(input_data):
    return input_data * 2

pipeline = Pipeline([node(process_data, "raw_data", "processed_data")])

Requires Python 3.10 or later; the data catalog and deployment integrations may require additional configuration or optional dependencies not bundled with the base package.

Verify before relying

• Whether the data catalog supports all advertised file formats and cloud providers out of the box or if additional plugins are required
• Performance characteristics and scalability limits for large pipelines or distributed execution
• Integration maturity with Argo, Prefect, Kubeflow, AWS Batch, and Databricks mentioned in the description

Package facts

Evidence: kedro-1.5.0-py3-none-any.whl

Keywords: pipelines, machine learning, data pipelines, data science, data engineering

Tags

More Application Frameworks packages