Now your whole team can write Argo Workflows.
dagweave is the visual editor for Argo Workflows. Build on a canvas, output YAML, and run it on your own clusters.
dagweave outputs a working Argo Workflows manifest. Commit it, apply it with kubectl, hand it to argo submit, or dispatch it from dagweave.
You already run Argo. Now the rest of your team can too.
Writing a Workflow means writing YAML. That keeps it with whoever already knows how.
Built around the Argo Workflows you already run.
Draw the graph or import one you already run. It compiles to Argo Workflows YAML as you edit, lands in your repo or dispatches to the cluster you pick.
The canvas
Pick a step from a searchable list and draw the dependencies. Retries, timeouts, node selectors and env are fields on the step.
Open what you already run
Paste a manifest, point at a GitHub repo, or read what is live on your cluster. Every field editable.
The step catalog
S3, Postgres, kubectl, Helm, docker builds, LLM calls. We write each WorkflowTemplate for you, version it, and keep its image current.
Every cluster you run, in one place
Enroll prod, staging and every region. Each trigger names the cluster it fires at, and you control who can send what where.
Start it on a schedule, a webhook or a push
Cron, webhooks, GitHub pushes and SNS topics, dispatched to the clusters you choose.
Watch it run on the graph you drew
Status streams back onto the canvas. Find the failed step where you drew it.
See where the time went
A bar per step against the clock. The wait before a pod started is drawn apart from the work it did.
Every run on the cluster
Crons, argo submit, whatever fires overnight. All of it in a section of its own, with nothing moved first.
Your YAML, your repo
A commit or a pull request against your own GitHub repo.
A second factor, and a record
One-time codes and passkeys on sign-in, and an audit log of who did what, whether it succeeded or was refused.
Groups, down to one workflow
Roles cover the workspace. Put people in groups and grant a group read or edit on a single workflow, template or connector.
Nothing left behind on your cluster
A run can get its own service account, its own namespace, or both. Whatever you give it goes when the run does.
Audit log
Every login, publish and dispatch recorded, filterable, and exportable as CSV.
Logs from your cluster
Every step's logs land on the step that produced them, wherever you keep them.
API and an MCP server
Runs, workflows and the audit log over HTTP, and an MCP server on the same routes. The assistant in your editor can pull a failed run itself.
Ask your assistant for a workflow
It drafts over MCP, we tell it what is broken, and you review a graph instead of a manifest you have to read line by line.
What building a workflow looks like.
Draw it from scratch, or edit one you already run
Pick each step from a searchable list and connect them. Or import a manifest that runs today and edit it visually on the canvas.
Read the YAML while you work
The compiled manifest sits in a pane beside the graph and updates as you edit.
Send it to your repo or a cluster
dagweave opens a pull request against your own repo. With a connector enrolled, it dispatches the same workflow and your cluster runs it.
Argo Workflows your whole team can use.
dagweave is in early access. Leave an email and we will bring you on as we open it up.