Node Types Reference

Schedule Trigger

Runs the workflow on a recurring cadence defined by a cron expression. Use this for periodic tasks like daily report generation, hourly inbox scanning, or weekly data sync.

Webhook Trigger

Starts the workflow when an external system sends an HTTP POST request to the webhook URL. The incoming request body is available as the trigger's output data.

Manual Trigger (Test Run)

Not a configurable trigger — every workflow can be executed manually via the Test Run button in the editor header. Useful during development and debugging.

Condition

Evaluates an expression and routes execution to one of two (or more) branches. Think of it as an if/then/else gate.

Field

The data path to evaluate (e.g., email.from).

Operator

Comparison operator — equals, not equals, contains, starts with, greater than, etc.

Value

The value to compare against.

Branches

True and False output handles. Each can connect to different downstream nodes.

You can chain multiple conditions with AND / OR logic for complex routing rules.

Loop

Iterates over an array and executes the connected downstream nodes once per item. The current item and index are available as variables inside the loop body.

Array path

The data path to the array (e.g., gmail.messages).

Item variable

Reference the current item as loop.item in downstream nodes.

Index variable

The zero-based index of the current iteration (loop.index).

Transform

Reshapes, renames, or computes new data from existing values. Use it to prepare data for the next node or to combine outputs from multiple upstream nodes.

• Rename fields — change 'first_name' to 'firstName'.
• Compute values — concatenate strings, perform arithmetic.
• Filter arrays — remove items that don't match criteria.
• Restructure JSON — flatten nested objects or nest flat objects.

HTTP Request

Makes an arbitrary HTTP request to any URL. Supports GET, POST, PUT, PATCH, and DELETE methods. Use this for APIs that don't have a dedicated integration.

AI Prompt

Sends a prompt to an LLM (Claude by default) and returns the response as the node's output. Use system and user message templates to structure the prompt.

Configuring Action Nodes

Every action node requires:

• A connection — the OAuth credentials for the target service. Select from the dropdown.
• A tool — the specific operation (e.g., gmail_send_email, slack_post_message).
• Parameters — the operation-specific inputs (recipient, channel, subject, etc.).

Parameters are dynamically generated based on the tool's schema. Required fields are marked. Optional fields are collapsed by default.

Integration Categories

See the full list with tool details in the Available Integrations reference.

Approval Gate

Pauses execution and creates an approval request that must be accepted or rejected by a team member before the workflow continues. See the Approval Gates guide for a full walkthrough.

Memory

Reads or writes persistent key-value data that survives across executions. Use it to track running totals, remember the last processed item, or store configuration that changes at runtime.

Sub-Workflow

Embeds another workflow as a single node. When the parent execution reaches this node, it starts a child execution of the referenced workflow, waits for it to complete, and passes the result back to the parent graph.