Skip to main content

Divides

A divide lets one Wrk Action handle many similar items at the same time instead of processing them one after another. Wrk splits the work automatically and runs the divided pieces in parallel — like handing each row in a spreadsheet to its own worker while you wait for the full set of results.

Divides are a form of parallel branching built into specific Wrk Actions. You do not draw separate branches on the canvas; the platform manages the split inside the Action.

What it looks like to you

On a Wrk Action card

Look for two signals when browsing or configuring a Wrk Action:

  1. Divide icon — shown next to the delivery methods on dividing Wrk Actions
  2. Description text — many dividing Actions state "This Wrk Action divides."

Divide icon on a Wrk Action card

On the canvas, a dividing Wrk Action still appears as a single card. The parallelism happens at execution time, not as extra connector lines you draw by hand.

During a run

When a dividing Wrk Action receives a list — spreadsheet rows, search hits, email attachments, line items from a document — Wrk fans that input out into parallel executions of the same step. Each slice completes on its own schedule, which is why large batches finish much faster than strict one-by-one processing.

Example: enriching many contacts in parallel

When to use divides

Good fit

  • Each item is independent — enriching contacts, downloading attachments, searching records per row
  • You need throughput more than strict ordering across items
  • The same Wrk Action logic applies to every element in a list

Poor fit

  • Items must be compared or merged before the next step (for example, picking the best of two strategies from combined results)
  • Downstream steps expect a single combined payload and you have not yet aggregated results

After parallel work completes, you can bring results back together — for example, with an Export Wrkflow data to a file Wrk Action.

Multiple divides in one Wrkflow

You can chain dividing Wrk Actions. Each divide creates its own parallel batch, so a Wrkflow that sends invitations, confirms RSVPs, and prepares welcome packets can divide at every step that supports it.

Divides vs loops

DividesLoops
Automatic inside qualifying Wrk ActionsExplicit loop + Continue Loop wiring
Parallel by defaultSequential unless you design concurrent patterns
Less canvas setupMore control over order, conditions, and exit

If you need to walk pages of an API in order, or stop when a counter reaches a limit, see Looping. Divides and loops often appear together in production Wrkflows.