Get on the list.
dagweave is the visual editor for Argo Workflows. It's pre-launch. Early access is the way in today. Leave an email and we'll get in touch as we bring people on.
Draw it or import it. Send it to your repo and your clusters.
The canvas and the YAML pane
Drag steps on and wire up the dependencies. Drill into nested sub-workflows as deep as they go. The manifest compiles next to the graph the whole time.
Your existing workflows, imported
From a GitHub repo, from a live cluster, or pasted in. Your whole file opens, retry strategies and artifact repository refs included. It goes back out as the same standard manifest, in your own repo.
Every cluster you run
One helm install per cluster, production-ready by default. Enroll as many clusters as you run. Authors dispatch to them without a kubeconfig of their own.
Straight into your GitHub repo
A direct commit or a pull request. What lands is a standard Argo Workflows manifest. It keeps working if you stop using dagweave.
Common questions.
What is the state of it today?
Pre-launch. Leave an email and we'll bring you on as we open it up.
Do you run my workflows?
No. dagweave is a builder, not a runtime. It writes Argo Workflows manifests and, once you enroll the connector, sends runs, WorkflowTemplates and CronWorkflows to your cluster, and reports what comes back. Argo Workflows executes all of it, on your own infrastructure.
Why Argo Workflows rather than something else?
CNCF-graduated, schedules on Kubernetes natively, and runs workloads that fail on other engines. dagweave adds the authoring layer on top and leaves the engine alone.
Do I have to choose between git and the connector?
No. The workflow lands in your repo as a manifest you own. dagweave dispatches that same file to whichever enrolled cluster you pick.
Am I locked in?
No. The output is a standard Argo Workflows manifest, sitting in your own repo, running on your own cluster. Stop using dagweave tomorrow, and every workflow carries on.
What does it install in my cluster?
A connector, from one helm chart: two replicas with leader election, a PodDisruptionBudget, one ClusterIP Service, no Ingress. A stock install also brings a DaemonSet for logs, so each step's logs land on the run itself. Log collection is optional.
What happens to my email?
We store your email to tell you when we open up. That's the whole use. No tracking, unsubscribe whenever.