Flow editor

Windmill's Flow editor allows you to build flows with a low-code builder.

Workflows are graphs (DAG) that represent job specifications. A workflow engine is a distributed system that takes a workflow and orchestrates it to completion on workers while respecting among others all the dependency constraints of each job. There exists a great variety of workflows and many software are domain-specific workflow engines and workflow specs (if you are a software engineer, you probably already wrote one without realizing it).

The goal of an orchestrator or workflow engine is to execute the steps on different workers that respect that graph's dependencies, and pass the required data defined in the steps' inputs while applying parallelism when possible.

In Windmill, each step is code with a main function in TypeScript, Python, Go, PHP, Bash or raw queries to databases. The result of a flow is the result of the last step executed, unless error was returned before or Early return is set.

Flows quickstart

4-minute presentation of our Flow editor.

Local development

Develop locally, push to git and deploy automatically to Windmill.

Architecture and data exchange

A workflow is a JSON serializable value in the OpenFlow format.

OpenFlow Spec

OpenFlow is an open standard for defining Flows.

Another approach is to write a program that defines the jobs and their dependencies, and then execute that program. This is known as workflows as code.

Workflows as code

Automate tasks and their flow with only code.

The Flow editor has the following features which are the subject of specific pages:

Triggering flows

Trigger scripts and flows on-demand, by schedule or on external events.

Testing flows

Iterate quickly and get control on your flow testing.

Flow editor components

Details on the flow editor's major components.

AI-generated flows

Generate flows from prompts.

Error handling in flows

There are four ways to handle errors in Windmill flows.

Branches

Split the execution of the flow based on a condition.

For loops

Iterate a series of tasks.

While loops

While loops execute a sequence of code indefinitely until the user cancels or a step set to Early stop stops.

Error handler

Configure a script to handle errors.

Trigger scripts

Trigger scripts are designed to pull data from an external source and return all of the new items since the last run, without resorting to external webhooks.

Retries

Re-try a step in case of error.

Concurrency limits

The Concurrency limit feature allows you to define concurrency limits for scripts and inline scripts within flows.

Custom timeout for step

If the execution takes longer than the time limit, the execution of the step will be interrupted.

Priority for steps

Prioritize a flow step in the execution queue.

Lifetime / Delete after use

The logs, arguments and results of this flow step will be completely deleted from Windmill once the flow is complete.

Cache for steps

Re-use a step's previous results.

Early stop / Break

Stop early a flow based on a step's result.

Early return

Define a node at which the flow will return at for sync endpoints. The rest of the flow will continue asynchronously.

Suspend & Approval / Prompts

Suspend a flow until specific event(s) are received, such as approvals or cancellations.

Sleep / Delays in flows

Executions within a flow can be suspended for a given time.

Step mocking

When a step is mocked, it will immediately return the mocked value without performing any computation.

Example of a flow in Windmill.