golem-parallel-workers-scala

name	golem-parallel-workers-scala
description	Fan out work to multiple parallel agents and collect results in a Scala Golem project. Use when the user asks about parallel execution, fan-out/fan-in, spawning child agents for parallel work, forking, or aggregating results from multiple agents.

Parallel Workers — Fan-Out / Fan-In (Scala)

Overview

Golem agents process invocations sequentially — a single agent cannot run work in parallel. To execute work concurrently, distribute it across multiple agent instances. This skill covers two approaches:

1. Child agents via codegen-generated XClient.get(id) — spawn separate agent instances, dispatch work, and collect results
2. HostApi.fork() — clone the current agent at the current execution point for lightweight parallel execution

Approach 1: Child Agent Fan-Out

Spawn child agents, dispatch work, and collect results using Future.sequence or Golem promises.

Basic Pattern with Future.sequence

import golem.*
import golem.runtime.annotations.{agentDefinition, agentImplementation}

import scala.concurrent.Future
import scala.scalajs.concurrent.JSExecutionContext.Implicits.queue

@agentDefinition()
trait Coordinator extends BaseAgent {
  class Id()
  def fanOut(items: List[String]): Future[List[String]]
}

@agentImplementation()
class CoordinatorImpl() extends Coordinator {
  override def fanOut(items: List[String]): Future[List[String]] = {
    // Spawn one child per item and call concurrently
    val futures = items.zipWithIndex.map { case (item, i) =>
      val child = WorkerClient.get(i)
      child.process(item)
    }

    // Wait for all children to finish
    Future.sequence(futures)
  }
}

@agentDefinition()
trait Worker extends BaseAgent {
  class Id(val id: Int)
  def process(data: String): Future[String]
}

@agentImplementation()
class WorkerImpl(private val id: Int) extends Worker {
  override def process(data: String): Future[String] = {
    Future.successful(s"processed-$data")
  }
}