Skip to main content

Actionflow Studio

Actionflow Studio is the visual flow editor. Build the graph on a canvas, then run the whole flow and inspect outputs, cost, and requirements.

Actionflow Studio

The Actionflow Studio is the flow editor at /actionflow-studio/{actionFlowId}. It is the same canvas you ship: start nodes, AI, integrations, control flow, and HITL. A run from the studio executes the full graph. Downstream nodes see real upstream outputs.

Actionflow Studio is not a node REPL. There is no isolated single-node runner, no Playground tab on the inspector, and no Test in Playground context-menu item. Edit the node, then run the flow.

Open it

  • Create a flow or open one from the dashboard. The editor is /actionflow-studio/{actionFlowId}.
  • Guest try mode: /actionflow-studio/try. Saving into an organization requires sign-in.

Layout

┌────┬──────────────────────────────────────────────┐
│Nav │  Top: flow name                               │
│ +  │                                              │
│hist│                    Canvas                   │
│art │                                              │
│set │                         inspector (on select)│
│hook│  Bottom: zoom, undo, Start, Run flow,          │
│use │          requirements, cost, outputs          │
└────┴──────────────────────────────────────────────┘

Left nav

ControlOpens
Back to ActionFlowsFlow list (unsaved-change guard)
Node CategoriesCatalog to add nodes
Flow Run HistoryRun history for this flow
ApprovalsHITL waiting on this flow
ArtifactsFiles this flow produced
SettingsFlow settings
WebhooksInbound flow webhook
UsageCredit usage for the flow

Try mode only shows back and Node Categories.

Canvas

Drop nodes, connect handles, pan and zoom. Right-click a node for Open, copy, duplicate, select all, or delete. Org members on the same flow see presence and cursors. You can follow another editor's cursor.

Node inspector (right)

Select a node to edit its inputs, model, credentials, and mapping. See Node settings. That is configuration for the step in the graph, not a separate test runner.

Bottom bar

  • Zoom, fit, lock, undo, redo
  • Start: saves the graph and starts a full run immediately
  • Run flow: opens Flow inputs for empty {nodeId}_{inputName} fields, then runs
  • Flow Requirements: missing credentials and similar blockers (shown when something is missing)
  • Cost estimate (when cost calculation is enabled for the org)
  • Run Output: per-node results and credit usage. Stop is available while a run is active

Studio and dashboard runs appear in history as triggerSource manual. The {nodeId}_{inputName} keys match API inputs and webhooks.

How to iterate

  1. Add a Start node (or Start Scheduled / Start Subflow).
  2. Add AI and action nodes. Wire execution and data handles. See Handles.
  3. Connect credentials from the org or the node inspector.
  4. Clear ActionFlow requirements.
  5. Click Start, or open Run flow if empty inputs still need values.
  6. Read outputs on the canvas and in Run Output. Adjust prompts or wiring. Run again.

Comparing models means changing the model on the AI node and running the same graph, then checking quality, tokens, and cost on that run. See Usage & Cost.

Studio vs run history

Actionflow StudioRun History
What it isThe editor plus a run you start from the canvasAudit of executions (studio, webhook, API, schedule, subflow)
What runsThe current graph (saved on Start)The graph that was saved for that run
Best forBuild, wire, and verify before you rely on webhooks or APIDebug a past run, filter, inspect cost

A studio run still consumes credits and still appears in history.

Collaboration

Org members who open the same flow share the room: avatars, cursors, and follow. Unsaved edits can block navigation away from Actionflow Studio.

Try Actionflow Studio

/actionflow-studio/try is a public canvas without an organization. Use it to try the editor, not production credentials, webhooks, or API keys.

On this page