Apache Airflow
airflow.apache.org
- Free plan
- macOS
- Linux
- Web
- API
- Self-hosted
What Apache Airflow is
Apache Airflow is a platform for authoring, scheduling, and monitoring workflows written in Python. It is designed for batch workflows with a defined start and end that run on a schedule. Its web application lets users manage and schedule workflows, trigger DAGs manually, inspect task status and logs, backfill runs, and rerun failed tasks. A modular architecture and message queue coordinate workers and support an arbitrary number of them. Built-in operators connect to Google Cloud Platform, Amazon Web Services, Microsoft Azure, and other services; the community maintains more than 80 separately installable provider packages. Providers can be upgraded or downgraded independently of Airflow core. Airflow can read connections, variables, and configuration from secret backends, but those integrations do not support writes. JWT authentication is used for its public REST API and internal Execution API. In Airflow 3 and later, sensitive credentials are masked in logs, the interface, and API outputs. Airflow is free to download and use under the Apache License.
Who it suits
Airflow suits teams running scheduled batch workflows that benefit from Python definitions and a web interface for monitoring and operations. It is intended for Linux-based deployment; macOS is for local development, while Windows deployment is unsupported except for WSL2 development.
Good to know, and what to watch
- Workflows are defined in Python.
- Web interface supports logs and task monitoring.
- Workers can scale through a modular architecture.
- More than 80 provider packages are available.
- Free to use under the Apache License.
- Secret-backend integrations do not support writes.
- Windows deployment is unsupported except WSL2 development.
- It is intended for scheduled batch workflows.
Plans
Every plan and priceApache Airflow
Free plan
Getting started
- Visit the Apache Airflow website
- Deploy Airflow in a self-hosted Linux environment
- Define workflows as Python code
- Install provider packages for required services
- Use the web application to schedule or trigger DAGs
- Monitor tasks and logs through the interface
Limits to check
The free Apache Airflow plan costs nothing and includes downloading and using the Apache License source code. Secret-backend integrations can read connections, variables, and configuration but cannot write them; Windows is not supported for deployment except WSL2 development.
The Club's review
Apache Airflow is a self-hosted platform for defining, scheduling, and monitoring workflows in Python. It is best suited to teams running scheduled batch work that want direct operational control. Its strongest case is extensible orchestration at no software cost; its Python requirement and deployment constraints narrow the fit.
Overview
Airflow puts workflow definitions in Python code and day-to-day oversight in a web application. That suits engineering teams who want workflows managed as code while operators can inspect task status and logs, start runs manually, backfill schedules, and rerun failures. It is designed for batch jobs with a clear start and finish on a schedule, not as a general answer for every kind of automation.
A modular architecture and message queue coordinate an arbitrary number of workers, offering a path to scale execution. The corresponding trade-off is operational ownership: Airflow is self-hosted, and official deployment support is for Linux-based environments. macOS is supported for local development; Windows is limited to WSL2 development and is not supported for deployment.
Key features
Workflow authoring and operations
Python is the workflow language, so Airflow favors teams able to define and maintain code rather than relying on a visual workflow builder; it has no visual builder. Its web interface supports manual DAG triggers, task-status and log inspection, backfills, and reruns of failed tasks. Automatic retries and execution observability help operators handle routine failures, while human approvals are supported for workflows that need a person involved.
Integrations and secrets
Plug-and-play operators connect workflows to Google Cloud Platform, Amazon Web Services, Microsoft Azure, and other third-party services. More than 80 community-maintained provider packages can be installed separately and upgraded or downgraded independently of Airflow core, which gives teams flexibility to manage integration changes without tying them to a core release.
Airflow can read connections, variables, and configuration from secret backends rather than its database, but those integrations are read-only: changes cannot be written back through them. JWT authentication is used for the public REST API and internal Execution API. In Airflow 3 and later, sensitive credentials are masked in logs, the interface, and API outputs.
Pricing
The Apache Airflow plan costs nothing. Download and use are free, and the source code is available under the Apache License. This makes it a practical option for teams that can operate their own deployment and want to avoid a software subscription; the cost advantage does not remove the work of running it.
There is no seat or usage cap stated for this plan. Project learning resources and community support are available through Slack and mailing lists. Teams that require a hosted service or vendor support should consider whether a self-hosted, community-supported project meets their operational needs.
Platforms
Airflow supports API and web access, with Linux and macOS among its platforms, and a self-hosted deployment model. For production deployment, official support is Linux-based; macOS is for local development, while Windows deployment is unsupported except for WSL2 development. That makes it a poor fit for organizations that require native Windows deployment.
Who it's for
Airflow is a strong candidate for engineering and data teams that run scheduled batch workflows, can author them in Python, and want a web interface for operational controls. Its worker architecture and provider ecosystem suit teams coordinating jobs across cloud services. Teams seeking visual authoring, a hosted service, or automation centered on event triggers rather than scheduled batch work should look elsewhere.
Pros and cons
- Pros: Free, Apache-licensed source code avoids a software fee for teams willing to self-host.
- Pros: Python definitions and operational controls keep workflow logic and run management close together.
- Pros: Separately managed provider packages and operators for major cloud platforms support integration flexibility.
- Cons: Self-hosting and Linux deployment support place deployment and operations responsibility on the adopting team.
- Cons: Python-based authoring and no visual builder make it less suitable for teams that need non-code workflow design.
- Cons: Secret-backend integrations cannot write changes, and Windows is not a supported deployment platform.
Alternatives
Kestra is another free-plan option, with a free trial and Windows among its supported platforms; consider it if Windows support is a deciding requirement. Flyte also offers a free plan and trial, and may suit teams comparing another self-hosted workflow platform.
DataKitchen TestGen offers a free tier with one user, one connection, and one project, alongside unlimited tables and data volume; it is worth considering when those caps fit the intended use. Apache NiFi is another free, open-source option with no commercial plan or usage limits stated on its official pages.
Apache Hop is a free open-source alternative for teams whose environment can meet its Java 21 requirement for release 2.19.0. Metaflow is free and open source, but production scaling requires infrastructure in the team's own cloud account or Kubernetes cluster.
Prefect may fit a small team that wants a managed serverless allowance: its free Hobby plan includes two users, one workspace, five deployments, 500 minutes per month, and seven days of run retention. Dagster is another orchestration framework with an open-source core and a free trial; consider it if evaluating a different framework.
For broader comparisons, browse DataOps Platforms, Workload Automation Software, or Workflow Orchestration Software.
Verdict
Choose Apache Airflow if your team can work in Python and wants free, extensible orchestration for scheduled batch jobs, backed by practical controls for monitoring and recovery. Look elsewhere if you need Windows deployment, a visual builder, or a hosted, event-first service; Airflow's strengths depend on accepting its self-hosted, code-centric model.
The verdict
Airflow provides workflow scheduling and operational controls for Python-based batch jobs, with integrations across cloud services. Its deployment and workflow fit matter: Windows is not supported for deployment, and secret backends are read-only.
Questions people ask
Is Apache Airflow free?
Yes. The Apache Airflow plan is free, with source code available under the Apache License.
Which platforms does Airflow support?
It supports API, Linux, macOS, self-hosted, and web environments. Linux is officially supported for deployment, while macOS is for local development; Windows deployment is unsupported except WSL2 development.
What language defines Airflow workflows?
Airflow workflows are defined in Python code.
Can Airflow connect to cloud services?
Yes. Built-in operators support Google Cloud Platform, Amazon Web Services, Microsoft Azure, and other services, with more than 80 separately installable provider packages.
Does Airflow have a visual workflow builder?
No. The listed capabilities include Python workflow definition and a web interface for managing workflows, but no visual builder.
What the maker publishes
- Authentication
- Airflow uses JWT authentication for its public REST API and internal Execution API. source
- Cloud integrations
- Built-in plug-and-play operators support Google Cloud Platform, Amazon Web Services, Microsoft Azure, and other third-party services. source
- Community support
- The project provides learning resources and community support through Slack and mailing lists. source
- Deployment platforms
- Airflow officially supports Linux-based deployment environments; macOS is supported for local development and Windows is unsupported for deployment except WSL2 development. source
- Founded
- 1999 source
- Headquarters
- Wilmington, Delaware, U.S. source
- Operations
- The web interface supports manually triggering DAGs, inspecting logs, monitoring task status, backfilling runs, and rerunning failed tasks. source
- Project history
- Airflow began in October 2014 at Airbnb and became an Apache Software Foundation top-level project in January 2019. source
- Provider releases
- Providers can be upgraded or downgraded independently of the Airflow core release. source
- Providers
- The community maintains more than 80 separately installable provider packages. source
- Purpose
- Apache Airflow is a platform for programmatically authoring, scheduling, and monitoring workflows. source
- Scalability
- Its modular architecture uses a message queue to orchestrate an arbitrary number of workers. source
- Secrets
- Airflow can read connections, variables, and configuration from secret backends instead of its database, but secret-backend integrations do not allow writes. source
- Sensitive data
- In Airflow 3 and later, sensitive credentials are masked in logs, the UI, and API outputs. source
- Web interface
- Its web application lets users monitor, schedule, and manage workflows and view task status and logs. source
- Workflow fit
- Airflow is intended for batch workflows with a clear start and end that run on a schedule. source
- Workflow language
- Airflow workflows are defined in Python code. source
Checked against the maker's pages on 1 Oct 2026.