Actionflow StudioCopy Anchor Link

Actionflow Studio is the visual editor at /actionflow-studio/{actionFlowId}. It is the same graph used by the other Flow entry points: start nodes, AI nodes, integrations, control nodes, storage, and HITL.

Before you runCopy Anchor Link

• Add exactly one supported start node and connect the executable graph from it.
• Configure required node inputs, named organization connections, and AI models.
• Save the graph. The Actionflow Studio run path explicitly saves the current graph before it calls the run API; the Flow inputs path also saves the submitted values.
• Fill empty required fields in Run flow when the run form exposes them. Their keys use the form {nodeId}_{inputName} convention.
• Review Flow Requirements for missing connections. Full validateFlow results can also report reachable-node input, model, credential, and Loop errors.

A passive flow is still editable. Actionflow Studio can run it while you build and test. The current API/Actionflow Studio run path does not enforce the webhook-style active-state check, so do not expect a 409 from that path. Webhook delivery rejects a passive flow with 409, and scheduled dispatch only claims active flows; activate the flow before relying on either production trigger. See Settings.

LayoutCopy Anchor Link

• Canvas: add nodes, connect execution/data handles, and move or arrange the graph.
• Inspector: select a node to edit its saved inputs, mappings, credentials, and model settings. The inspector is configuration, not a separate test runner.
• Left navigation: open run history, approvals, artifacts, settings, webhooks, and usage for the flow.
• Bottom bar: run the full graph, open Flow inputs, inspect requirements, view the optional cost estimate, and stop an active run.

Run controlsCopy Anchor Link

Actionflow Studio and dashboard runs use manual as their trigger source. API-key runs use api by default. A webhook or schedule has its own source and response contract.

Live output versus run historyCopy Anchor Link

The active Actionflow Studio output can show current run metadata, progress, and node results while the run is moving. When the run reaches a terminal state, the app can load the persisted run-history payload for the detail view.

These are not guaranteed to be byte-for-byte identical:

• History is a filtered view. The app's history read path sanitizes data and redacts sensitive-looking keys.
• Provider responses and large loop results can be represented by a summary, storage key, or truncation marker.
• A missing path can produce an empty value even when the reference is visible in the editor.

Use live output for immediate debugging and Run history for the persisted audit view. See Execution model for the 5 MB remote HTTP response limit and the current loop-history result cap.

CreditsCopy Anchor Link

Starting a run consumes organization credits. The runtime reserves a credit hold at admission, then settles the actual amount from executed node and AI usage. A failed run can retain billable usage from the prefix that completed; a cancelled or failed run with no billable usage can release its hold. The Actionflow Studio estimate is not a bill and can differ from the settled amount. See Usage and Organization permissions.

Collaboration and try modeCopy Anchor Link

Organization members editing the same flow can see presence and cursors. Unsaved edits can block navigation away from Actionflow Studio. /actionflow-studio/try is a public editor mode without an organization; it is for trying the editor, not production credentials, webhooks, or organization billing.

RelatedCopy Anchor Link

• Execution model
• Troubleshooting
• Node settings
• Node handles
• Run history
• Flow requirements
• Credential management
• Artifacts
• Settings