Open Source Workflow Automation Tools (2026): 9 to Self-Host

A familiar story: someone builds a set of automations for their business on a self-hosted tool, it works beautifully, and then they want to offer the same set-up to clients. That's when they discover the licence they never read has opinions about it.

So this page starts with licences, then groups nine self-hostable tools by who'll be building the workflows. They're in no particular order and nothing here is ranked. Most of them, outside the visual builders, need someone who's happy writing code.

Fair-code, open core and plain open source

The best-known self-hosted builder, n8n, describes itself as fair-code. Its code sits under the Sustainable Use License, which makes it source-available rather than open source in the strict sense the Open Source Initiative uses, and files marked as enterprise sit under a separate commercial licence. You can read the code and self-host it, but the licence limits some uses.

Two others are open core. Activepieces is MIT for most of its code, with an enterprise folder under a commercial licence. Windmill is mostly AGPLv3, with Apache 2.0 client libraries and proprietary enterprise-flagged code; the AGPL part carries obligations if you offer a modified copy to others as a service.

The other six use Apache 2.0 or MIT, which are permissive. Before you build anything important, check two things: which features live in the commercial folder, and whether your use, an internal tool or a service you resell, fits the licence.

Visual builders for people who don't write code

These give you a canvas where you connect steps without programming. Two of them are the licence cases above: a fair-code node editor and a Zapier-style open core builder. The third, Node-RED, is a low-code flow editor in the browser from the OpenJS Foundation, under Apache 2.0, which has long been popular for wiring devices and APIs together.

Code-first platforms for developers

Here workflows are written in code, which suits teams who want version control and testing. The open core platform above sits here too, pitching itself as a self-hostable home for scripts, flows and internal apps. Trigger.dev focuses on background jobs and AI agents, with a self-hosting guide. Temporal is a durable execution engine under MIT, built so long-running processes survive crashes and restarts.

Schedulers and orchestrators for data teams

These are for data and infrastructure teams moving and transforming data on a schedule. Apache Airflow defines workflows as Python code and is an Apache Software Foundation project. Prefect turns Python scripts into production pipelines, and its open source version is the one you run yourself. Kestra takes a declarative, event-driven approach to data, AI and infrastructure workflows.

All nine tools compared

Tool	Who builds	Good for
Activepieces	Visual builder, open core	No-code automations with TypeScript integration pieces
Apache Airflow	Python scheduler	Authoring, scheduling and monitoring workflows as Python code
Kestra	Declarative orchestration	Event-driven data, AI and infrastructure workflows
n8n	Visual builder, fair-code	Self-hosted workflows and AI agents on a node canvas
Node-RED	Flow editor, event-driven	Wiring devices, APIs and events in a browser editor
Prefect	Python pipeline framework	Turning Python scripts into production data pipelines
Temporal	Durable execution engine	Applications that keep running through failures
Trigger.dev	Background jobs in code	Background jobs, AI agents and workflows written in code
Windmill	Code-first, open core	Scripts, flows, internal apps and AI agents

Advice for picking an automation tool you'll still be happy with

If nobody on your team writes code, a visual builder is the realistic choice, and that means reading its licence before anything else. If you only automate your own business, fair-code and open core licences rarely get in the way. If you're an agency or consultant planning to run automations for clients, look hard at the terms, or pick a permissively licensed tool and accept that it may need more technical set-up.

If you've got developers, a code-first platform gives you workflows you can review and test like any other code. Data teams will already know whether they need a scheduler. For hosted tools that need no server at all, our guide to workflow automation tools covers that side.

The nine tools, one line each

Activepieces

Best for: Non-developers wanting a no-code automation builder pitched as an open source replacement for Zapier, with integrations written as TypeScript pieces.

Apache Airflow

Best for: Data engineering teams who want to author, schedule and monitor workflows written as Python code, under an Apache Software Foundation project.

Kestra

Best for: Platform and data teams orchestrating event-driven workflows for data, AI and infrastructure, described declaratively rather than in a visual canvas, the project says.

n8n

Best for: Teams wanting a self-hosted workflow and AI agent platform built around a visual node editor, who are comfortable with a fair-code licence.

Node-RED

Best for: Makers and engineers wiring together devices, APIs and event-driven applications in a low-code browser flow editor from the OpenJS Foundation.

Prefect

Best for: Python developers turning scripts into production data pipelines, who run the open source version themselves, with a separate hosted product on offer.

Temporal

Best for: Engineering teams building applications that must keep running through crashes and failures, using a durable execution platform under MIT.

Trigger.dev

Best for: Developers building and deploying background jobs, AI agents and workflows in code, with a self-hosting guide alongside its hosted product.

Windmill

Best for: Developer teams wanting a code-first, self-hostable platform for workflows, internal apps and AI agents, while noting some enterprise code is proprietary.

Internal tool or client service: where the licence bites

Running a self-hosted automation tool for your own team is the use every licence here allows. The questions start when you host it for others. A source-available licence may restrict offering the tool itself as a service. AGPL code asks you to publish your changes if you offer a modified copy over a network. Enterprise folders may hold the features, such as single sign-on or audit logs, that client work tends to need.

None of that rules a tool out. It just means the licence belongs in your plan on day one, next to the server and the backups. And whichever tool you pick, someone has to update it, monitor failed runs and keep credentials safe, because self-hosted automation is only as reliable as the person looking after it. If you're weighing the fair-code builder against a hosted rival, our comparison of a self-hosted and a hosted automation builder goes further.

Do you make a workflow automation tool? Tool makers can ask to be considered for a featured place on this list.

Want your tool on this list? Featured spots and reviews here are paid partnerships. Here's how to partner with me.

Frequently asked questions

Is fair-code the same as open source?

Not in the strict sense. The best-known fair-code builder's own README uses that term, and its code sits under the Sustainable Use License, which makes it source-available. You can read and self-host the code, but the licence limits some commercial uses.

What does open core mean for an automation tool?

It means most of the code is under an open licence while some features, usually enterprise ones, sit in a separate folder under a commercial licence. Check which side the features you need are on.

Can I resell automations built on a self-hosted tool?

It depends on the licence. Permissive licences such as Apache 2.0 and MIT allow most uses. Source-available and AGPL licences can restrict or add conditions to offering the tool as a service, so read the terms first.

Do I need to code to use open source workflow automation?

For the visual builders, mostly not. For the code-first platforms and data schedulers, yes: workflows are written in Python, TypeScript or similar, so they suit developers and data teams.

What's the difference between a workflow builder and an orchestrator?

A workflow builder connects apps and triggers, often visually, for business tasks. An orchestrator schedules and monitors larger jobs, such as data pipelines, and usually expects workflows defined in code.

What does self-hosting an automation tool involve?

You run the server, keep the software updated, back up your workflows and credentials, and watch for failed runs. The software is free to use under its licence, but the upkeep is yours.