Automation & Workflows

Connect apps and automate workflows with self-hosted alternatives to Zapier and Make.

72 self-hosted apps · 199 comparisons

All automation & workflows apps

Last reviewed Aug 26, 2026 · 438 words

With 202,435 GitHub stars, n8n has effectively won the self-hosted Zapier race, and for most people the category decision is already made. The interesting question is the one nobody asks: whether you need a workflow platform at all, or a smaller tool that does your one job better with a tenth of the attention.

How to choose an automation stack

The first axis is breadth versus purpose. n8n connects hundreds of services in a visual node editor with code steps when you need them; it is the right shape when your automations span many apps and keep multiplying. But if the actual job is "tell me when this page changes", changedetection.io does exactly that in a 256 MB container with visual diffs and a pile of notification channels, no workflow editor required. The second axis is who writes the automations. Visual editors suit mixed-skill households and teams; Apache Airflow defines everything as Python DAGs, which is exactly what a data team wants and exactly what everyone else should avoid. The third axis is licensing: n8n's Sustainable Use License is fair-code, not OSI-approved open source. Self-hosting it for your own business is fine; building a product on it needs a careful read.

The Airflow mistake

Airflow's name recognition pulls homelabbers toward a 4 GB, Hard-difficulty orchestrator built for data-engineering pipelines with scheduler, webserver, and worker processes to babysit. If your workflows are "when X happens, do Y" rather than "rebuild this dependency graph of batch jobs nightly", Airflow is the wrong tool at any scale you will reach at home. The same logic applies in miniature to Huginn: charming, MIT-licensed, and technical enough that n8n usually gets the same result in a fifth of the time.

Where to start

n8n for anyone whose automations touch 3 or more services: the integration library is unmatched in self-hosting and a single Docker container gets you going in 512 MB. changedetection.io for price watching, restock alerts, and content monitoring, even alongside n8n, because a dedicated watcher beats a workflow you have to build. Airflow only if you are orchestrating actual data pipelines and know why you want DAGs. If your automations lean toward smart-home events and MQTT rather than web APIs, read the n8n vs Node-RED comparison before committing.

For a first-timer the default is n8n, self-hosted with no execution caps, with one warning from experience: workflows grow gnarly quietly, so name them well and export backups before the tenth one exists.

Curated picks

Automation & Workflows comparisons

199 head-to-head comparisons in this category.

See all comparisons →