Getting started

This page takes you from an empty canvas to real Argo Workflows YAML you own. No cluster and no execution required to get there. When you want dagweave to run the workflow for you, the last section covers that too.

dagweave is the visual editor for Argo Workflows. You build a pipeline as a graph and it writes the YAML for you, so you do not have to memorise the Argo Workflows schema to get a working manifest.

Two ways to use dagweave

dagweave splits into two roles, and which one you are decides where you start:

• Workflow authors build workflows on the canvas. This is most people. You do not need to touch Kubernetes or install anything to author a workflow and take its YAML.
• Cluster operators install the connector once so authors can run their workflows on your Argo Workflows cluster. You only do this if your team wants dagweave to submit runs.

If you are here to build workflows, keep reading. If you are here to connect a cluster, jump to Installing the connector.

Open the builder

The builder runs in your browser. There is nothing to install and no local dependency to set up. Sign in and you land on an empty canvas.

If you want to run a workflow on your own machine instead of a cluster, that is a separate, free binary: see Dagweave Local.

Build your first workflow

1. Pick your first step from the searchable list. Each node is one step in your pipeline.
2. Set its container image and the command it runs.
3. Add a second node and draw an edge from the first to it. The edge is a dependency: the second step waits for the first. You can also wire an output of one step into an input of the next, so data flows down the graph.
4. Open Settings and pick the kind you want to produce: WorkflowTemplate, Workflow, or CronWorkflow.
5. Watch the YAML preview update as you go. When it looks right, copy it or download the file.

That YAML is real Argo Workflows, and it is yours. Apply it the same way you would a manifest you wrote by hand:

kubectl apply -f my-pipeline.yaml
# or, for a one-off run
argo submit my-pipeline.yaml

This is the free path start to finish: build visually, get real YAML, take it to your own cluster. You never have to hand dagweave your cluster to get value from it.

The node types

dagweave ships a set of core nodes that map to common pipeline steps. You wire them together instead of writing container templates by hand:

• run-script runs an inline script in a container.
• run-container runs a prebuilt image with your command and args.
• http-request calls an HTTP endpoint.
• k8s-resource creates or patches a Kubernetes resource.
• approval pauses the run and waits for a human to resume it.

Every node has a config form driven by its schema, so you fill in fields instead of guessing at YAML keys. You can also add global parameters and reference them with {{workflow.parameters.x}} in any field, and repeat a node over a list to fan out work.

Git checkout and commit-push are core nodes, and the catalog also carries versioned copies of them. The core ones are what you wire: catalog items have no ports, and these two are the only nodes that move a working tree between steps, so without them in the palette an artifact input cannot be satisfied from an empty canvas. See the node reference for every type.

Run it on your cluster

Copying YAML is enough if you drive Argo Workflows yourself. If you want dagweave to submit runs and show you the results, a cluster operator installs the connector on your cluster. It opens a single outbound channel to dagweave, so your cluster never accepts inbound traffic. See Installing the connector.

Connecting a cluster is a paid feature, so Add a connector is refused for a team whose plan does not include it. A team owner can see where the team stands under Plans and billing, which is worth doing before an operator books time to install one.

Once a cluster is connected you can attach triggers to a workflow (a button, a cron schedule, a webhook, or an event) and watch each run in the executions view with logs, metrics, and re-run.

Next steps

• Core concepts explains how the canvas compiles, how you wire data between steps, and how import round-trips your existing YAML.
• Installing the connector connects your Argo Workflows to dagweave.
• Triggers covers manual, cron, webhook, and event runs.
• Executions covers watching runs and reading logs.