Installing and Running Locally

Use a venv and the constraints file

Airflow has many dependencies, so install it in a virtual environment with the project's constraints file for your Airflow and Python versions; that pins compatible versions and avoids dependency conflicts. For learning, airflow standalone initialises the database, creates an admin user and starts all components on one machine. For anything beyond learning, use Docker Compose, the official Helm chart or a managed service (Amazon MWAA, Google Cloud Composer, Astronomer). Set AIRFLOW_HOME to choose where config, logs and the DAGs folder live.

Install with constraints

Replace the version numbers with the release you want. Airflow 3.x requires a supported Python 3 version; check the install docs.

python3 -m venv af-venv && source af-venv/bin/activate
AIRFLOW_VERSION=3.0.6
PYTHON_VERSION="$(python -c 'import sys; print(f"{sys.version_info.major}.{sys.version_info.minor}")')"
pip install "apache-airflow==${AIRFLOW_VERSION}" \
  --constraint "https://raw.githubusercontent.com/apache/airflow/constraints-${AIRFLOW_VERSION}/constraints-${PYTHON_VERSION}.txt"

export AIRFLOW_HOME=~/airflow
airflow standalone          # learning only: DB + admin user + all components

Do not use standalone in production

It uses a local SQLite database and a single process, which is fine for learning but not for concurrent, reliable production use.

Quick check: Why install Airflow with a constraints file?

• It is required to open the UI
• It makes Airflow free
• It removes the scheduler
• It pins compatible dependency versions and avoids conflicts