Split Step

Loop over a list and run a subflow for each item.

What it does

• Iterates through an array and executes a nested set of steps for every item.
• Can run sequentially or in parallel and optionally keep going even if some items fail.

When to use it

• To process collections like order line items, rows from a CSV, pages of API results, or files in a batch.
• To fan out work that repeats the same logic per element.

How to configure

• Point value to the list (e.g., data.items, steps.fetch.outputs.records).
• Build the subgraph: add the steps that should run for each item.
• Optionally set parallelism (0 = no limit), choose sequential if order matters, and decide whether to stop on first failure.
• Define input/output schemas when you want clearer contracts for each iteration’s data.

Typical patterns

• For each line item, reserve inventory, then aggregate results.
• For each product, normalize fields and upsert to a storefront.
• For each record, call an external API and collect responses for the Finish step.