Skip to content

Guides

Task-oriented guides for common workflows. If you're new, start with Getting Started. If you know what you want to do, jump directly to the relevant guide.

Getting Started

GuideWhat you'll learn
InstallationInstall truffle, spawn, and optional tools
Your First InstanceEnd-to-end: find, launch, connect, terminate

Compute

GuideWhat you'll learn
GPU Training JobsLaunch a GPU instance, run a training job, terminate on completion
Jupyter NotebooksRStudio or Jupyter with process-aware idle detection
Spot InstancesUse Spot for up to 90% cost savings, handle interruptions gracefully

Automation & Control

GuideWhat you'll learn
Slack SetupConnect Slack, use /spore commands, get lifecycle DMs
Teams SetupConnect Microsoft Teams
AI Assistant (MCP)Control compute from Claude Desktop or Cursor
Lifecycle NotificationsConfigure what events trigger notifications and where they go

Advanced

GuideWhat you'll learn
Parameter SweepsRun the same job across many parameter combinations in parallel
MPI ClustersLaunch multi-node clusters for distributed workloads
PipelinesChain jobs so each stage launches the next
PluginsExtend instances with pre-built or custom plugins at launch time
Job ArraysLaunch and manage groups of related instances

Workflow Engines

GuideWhat you'll learn
Workflow EnginesRun Nextflow, WDL, CWL, Snakemake, or Airflow with spawn as the execution backend — each task on an ephemeral instance
Nextflow (nf-spawn)Dispatch each Nextflow process to its own ephemeral EC2 instance