Block: Parallel Block — Zelaxy Docs

Container block that executes child workflow blocks in parallel. Supports both count-based parallel execution and collection-based distribution. Child blocks placed inside the parallel container will be executed concurrently. Use it whenever independent tasks can run at the same time and you want to reduce overall workflow execution time.

Overview

Property	Value
Type	`parallel`
Category	`blocks`
Color	`#FEE12B`

When to Use

Run independent API calls simultaneously instead of sequentially
Process every item in a list concurrently (e.g., enrich each row of a dataset in parallel)
Query multiple data sources at the same time and collect all results
Generate multiple AI outputs in parallel (e.g., the same prompt sent to several providers)
Speed up workflows where tasks do not depend on each other's output
Control concurrency limits to avoid hitting external rate limits

Configuration

Parallel Type

Dropdown — selects the mode of parallel execution.

Option	id	Description
Parallel Count	`count`	Runs child blocks a fixed number of times concurrently
Parallel Each (Collection)	`collection`	Distributes items from an array or object across parallel executions

Parallel Executions

Short input — only shown when Parallel Type is count.

id: count
Placeholder: 3
Specifies how many concurrent executions to launch.

Distribution Items

Long input — only shown when Parallel Type is collection.

id: collection
Placeholder: ['item1', 'item2', 'item3'] or {{previous_block.output}}
Array or object to distribute across parallel executions. Use {{block_id.output}} to reference another block's output.

Max Concurrency

Slider — always visible.

id: maxConcurrency
Range: 1 – 20
Maximum number of parallel executions allowed to run at the same time. Useful for staying within external API rate limits.

Wait for All

Switch — always visible.

id: waitForAll
When enabled, the block waits for every parallel execution to finish before the workflow continues downstream.

Stop on Error

Switch — always visible.

id: stopOnError
When enabled, all in-flight parallel executions are cancelled as soon as any single execution fails.

Timeout (seconds)

Short input — always visible.

id: timeout
Placeholder: 300
Maximum wall-clock time (in seconds) to wait for all executions to complete before the block times out.

Inputs & Outputs

Inputs

collection (json) — Collection of items to distribute across parallel executions
count (number) — Number of parallel executions

Outputs

executions (number) — Total number of parallel executions completed
results (json) — Array of results from each parallel execution
executionTime (number) — Total execution time in milliseconds
status (string) — Parallel execution status (completed, error, or cancelled)
currentExecution (number) — Current execution number (available during execution)
currentItem (json) — Current item being processed (for collection-based parallel execution)

Tools

This block uses no external tools (tools.access is empty). All execution logic is handled internally by the parallel executor handler.

YAML Example

parallel_1:
  type: parallel
  name: "Parallel"
  inputs:
    parallelType: "collection"
    collection: "{{fetch_data_1.results}}"
    maxConcurrency: 5
    waitForAll: true
    stopOnError: false
    timeout: "120"
  connections:
    outgoing:
      - target: next-block-id

Parallel Block