Every operations team has the same artifact: a process doc that describes how orders are handled, how leads are routed, how invoices get chased — and a reality that drifts further from it every quarter. The doc can't execute. FluidGrids makes the diagram itself the thing that runs.
Open the builder, click New, and you get a blank canvas with a node palette. Drag an API trigger on, then the steps: a set node to shape data, an if node to branch, a Slack node to notify. Wiring is drag-to-connect, and the validator rejects connections that don't make structural sense — dangling branches and invalid edges are caught before you ever hit save.
The canvas is genuinely a builder, not a picture. Drop a new node onto an existing edge and it splices itself in. Delete a node and the graph heals. Auto-layout keeps a growing workflow readable. Each node carries its own configuration, so the diagram and the behavior are never two separate sources of truth.
When you save, FluidGrids persists the entire graph as a versioned workflow definition under a stable workflow key. Hard-refresh the page and the identical graph re-renders — nodes, edges, positions, and configuration all round-trip through the backend. Every subsequent save can mint a new version, so iteration never destroys what came before.
That definition is the foundation everything else builds on: triggers attach to it, background workers execute it, run history pins to the exact version that ran, and the same canvas ships inside the embeddable SDK. One drawing, one source of truth, from first sketch to production.
Do it yourself
Turn a process doc into a versioned, executable workflow on the drag-and-drop canvas — drag a trigger and a few action nodes, wire them, configure each, and save your first version.
Open the My Workflows page and click New to start a blank workflow.
You should see: A blank canvas opens in the builder with the node palette down the side.
Ready to make this your story?
