spawn command reference

Global flags

All spawn commands inherit these persistent flags.

Flag	Short	Type	Default	Description
--output	-o	string	table	Output format: table, json
--verbose	-v	bool	false	Enable verbose output
--no-color		bool	false	Disable colorized output
--lang		string	(system)	Language for output: en, es, fr, de, ja, pt
--no-emoji		bool	false	Disable emoji in output
--accessibility		bool	false	Enable accessibility mode (implies --no-emoji)

---

spawn launch

Launch an EC2 instance with automatic lifecycle management.

spawn launch <name> [flags]

<name> is required. It sets the EC2 Name tag, the DNS hostname (<name>.<account>.spore.host), and the instance's own hostname. Names must be unique within your account.

Examples:

# Basic launch — 1h idle timeout auto-applied if no --ttl
spawn launch my-job --instance-type c6a.4xlarge

# GPU training with TTL, Spot, and auto-terminate on completion
spawn launch train --instance-type p4d.24xlarge --spot --ttl 48h --on-complete terminate

# Job array (3 identical instances)
spawn launch workers --instance-type m7i.xlarge --count 3 --job-array-name workers

# Parameter sweep from YAML file
spawn launch sweep --instance-type c6i.2xlarge --param-file params.yaml --cartesian

# MPI cluster with EFA
spawn launch mpi-job --instance-type hpc7i.48xlarge --count 4 --mpi --efa

# With pre-stop S3 sync before lifecycle termination
spawn launch analysis --instance-type r7i.4xlarge --ttl 8h \
  --pre-stop "aws s3 sync /results s3://bucket/run-001"

# Larger root volume
spawn launch big-disk --instance-type m7i.large --volume-size 200

JSON output: With -o json, the TUI progress display is suppressed and spawn launch emits a parseable JSON array of launched instances, each with instance_id, name, instance_type, region, public_ip, state, and dns. Capture the ID in scripts with:

id=$(spawn launch my-job --instance-type c6a.large -y -o json | jq -r '.[0].instance_id')

Instance configuration

Flag	Type	Default	Description
--instance-type	string		EC2 instance type
--region	string	(config default)	AWS region
--az	string		Specific availability zone
--ami	string	(auto-detects AL2023)	AMI ID. Use auto (or omit) to auto-detect the latest AL2023 AMI
--volume-size	int	(AMI default)	Root EBS volume size in GiB. Use for AMI-baking or large datasets that exceed the default root volume
--vpc	string		VPC ID (auto-creates if unset)
--subnet	string		Subnet ID
--security-group	string		Security group ID
--key-pair	string		SSH key pair name

Capacity

Flag	Type	Default	Description
--spot	bool	false	Launch as Spot instance
--spot-max-price	string		Maximum Spot price (e.g., 0.50)
--use-reservation	bool	false	Use On-Demand Capacity Reservation
--reservation-id	string		Capacity Reservation / Capacity Block ID to launch into (cr-…). The instance must be in the reservation's AZ (set --az)
--capacity-block	bool	false	The --reservation-id is a Capacity Block for ML (sets MarketType=capacity-block); mutually exclusive with --spot

Lifecycle

Flag	Type	Default	Description
--ttl	string		Auto-terminate after duration (e.g., 8h, 2d). If neither --ttl nor --idle-timeout are set, a 1h idle timeout is applied automatically.
--idle-timeout	string	1h	Auto-terminate if no activity for this duration. Disabled if --ttl is set unless also explicitly specified.
--no-timeout	bool	false	Disable all automatic timeout (creates zombie risk — use with caution)
--hibernate	bool	false	Enable hibernation support
--hibernate-on-idle	bool	false	Hibernate instead of terminate when idle
--on-complete	string		Action when workload signals completion: terminate, stop, hibernate
--completion-file	string	/tmp/SPAWN_COMPLETE	File path that signals completion
--completion-delay	string	30s	Grace period after completion signal
--pre-stop	string		Shell command to run before any lifecycle-triggered stop/terminate
--pre-stop-timeout	string	5m (Spot: 90s)	Max time for --pre-stop command
--session-timeout	string	30m	Auto-logout idle SSH sessions (0 to disable)

Notifications and idle detection

Flag	Type	Default	Description
--slack-workspace	string		Slack workspace ID for lifecycle notifications (e.g., T03NE3GTY)
--active-ports	string		Comma-separated TCP ports to monitor — won't idle-terminate while any are active (e.g., 8787 for RStudio, 8787,8888 for RStudio+Jupyter)
--active-processes	string		Comma-separated process names to monitor — won't idle-terminate while any are running (e.g., rsession, rsession,jupyter)

Networking

Flag	Type	Default	Description
--dns	string		Override DNS name if different from --name
--dns-domain	string		Custom DNS domain
--dns-api-endpoint	string		Custom DNS API endpoint

Startup

Flag	Type	Default	Description
--user-data	string		Inline user-data script or @/path/to/file
--user-data-file	string		Path to user-data script file
--plugin	strings		Plugin to install at launch (repeatable, e.g., --plugin my-plugin@1.0)
--config	string		Launch config YAML file (supports plugins: list)
--command	string		Command to run on all instances after setup

Shared storage

Flag	Type	Default	Description
--efs-id	string		EFS filesystem ID to mount (e.g., fs-abc123)
--efs-mount-point	string	/efs	EFS mount point
--efs-profile	string	general	EFS performance profile: general, max-io, max-throughput, burst
--efs-mount-options	string		Custom EFS mount options (overrides profile)
--fsx-id	string		Existing FSx Lustre filesystem ID to mount
--fsx-create	bool	false	Create new FSx Lustre (PERSISTENT_2, Lustre 2.15, compatible with AL2023) with S3 backing via Data Repository Association. Writes spawn:fsx-id, spawn:fsx-mount-name, spawn:fsx-mount-point tags on all instances.
--fsx-id	string		Existing FSx filesystem ID. Writes spawn:fsx-id, spawn:fsx-mount-name, spawn:fsx-mount-point tags so boot scripts can auto-mount without hardcoding the ID.
--fsx-recall	string		Recall FSx by CloudFormation stack name
--fsx-storage-capacity	int	1200	FSx storage in GB (1200, 2400, or multiples of 2400)
--fsx-throughput	int	125	PERSISTENT_2 throughput in MB/s/TiB — valid values: 125, 250, 500, 1000
--fsx-s3-bucket	string		S3 bucket for FSx (required with --fsx-create)
--fsx-import-path	string		S3 import path
--fsx-export-path	string		S3 export path
--fsx-mount-point	string	/fsx	FSx mount point

Job arrays

Flag	Type	Default	Description
--count	int	1	Number of instances to launch
--job-array-name	string		Job array name (required when --count > 1)
--instance-names	string		Instance name template (e.g., worker-{index})

MPI clusters

Flag	Type	Default	Description
--mpi	bool	false	Enable MPI cluster setup (requires --count > 1)
--efa	bool	false	Enable Elastic Fabric Adapter for ultra-low-latency MPI
--mpi-processes-per-node	int	(vCPU count)	MPI processes per node
--mpi-command	string		Command to run via mpirun
--skip-mpi-install	bool	false	Skip MPI installation (for AMIs with MPI pre-installed)
--placement-group	string		Use an existing AWS Placement Group by name
--auto-placement-group	bool	true	Auto-create a cluster placement group named spawn-mpi-<job-array-name> and wait for it to become available before launching instances

Parameter sweeps

Flag	Type	Default	Description
--param-file	string		Path to parameter file (JSON, YAML, or CSV)
--params	string		Inline JSON parameters
--cartesian	bool	false	Generate cartesian product of parameter lists
--sweep-name	string	(auto)	Human-readable sweep identifier
--max-concurrent	int	0	Max simultaneous instances (0 = unlimited)
--max-concurrent-per-region	int	0	Max simultaneous instances per region
--launch-delay	string	0s	Delay between instance launches
--detach	bool	false	Run sweep orchestration in Lambda (auto-enabled for sweeps)
--no-detach	bool	false	Disable auto-detach (requires --ttl or --idle-timeout)
--budget	float	0	Budget limit in dollars (0 = no limit)
--cost-limit	float	0	Terminate when spend reaches this amount (0 = disabled)
--estimate-only	bool	false	Show cost estimate and exit without launching
--yes / -y	bool	false	Auto-approve cost estimate
--mode	string	balanced	Distribution mode: balanced or opportunistic

Region constraints (for sweeps)

Flag	Type	Default	Description
--regions-include	strings		Only use these regions (wildcards: us-*)
--regions-exclude	strings		Exclude these regions (wildcards: eu-*)
--regions-geographic	strings		Geographic filter: us, eu, ap, north-america
--proximity-from	string		Prefer regions close to this region
--cost-tier	string		Prefer cost tier: low, standard, premium

IAM

Flag	Type	Default	Description
--iam-role	string		IAM role name (creates if doesn't exist)
--iam-policy	strings		Service-level policies (e.g., s3:ReadOnly)
--iam-managed-policies	strings		AWS managed policy ARNs
--iam-policy-file	string		Custom IAM policy JSON file
--iam-trust-services	strings	ec2	Services that can assume the role

Compliance

Flag	Type	Default	Description
--nist-800-171	bool	false	Enable NIST 800-171 Rev 3 compliance mode
--nist-800-53	string		Enable NIST 800-53 compliance level: low, moderate, high
--compliance-strict	bool	false	Fail on compliance warnings (default: warn only)

Workflow integration

Flag	Type	Default	Description
--output-id	string		Write instance/sweep ID to file for scripting
--wait	bool	false	Wait for sweep/launch to complete
--wait-timeout	string	0	Timeout for --wait (0 = no timeout)
--team	string		Team ID for team-shared instance access
--interactive	bool	false	Force interactive wizard
--quiet	bool	false	Minimal output

---

spawn list

List running instances.

spawn list [flags]
spawn ls [flags]

Groups output by parameter sweeps, job arrays, and standalone instances. Shows columns: Instance ID, Name, Type, State, IAM Role, AZ, Age, TTL, Public IP, Spot.

Examples:

spawn list
spawn ls --regions us-east-1,us-west-2
spawn list --state running --instance-family m7i
spawn list --sweep-name my-sweep
spawn list --output json

Flags:

Flag	Short	Type	Default	Description
--region		string	(all)	Filter by a single region
--regions	-r	strings	(all)	Filter by multiple regions (comma-separated)
--az		string		Filter by availability zone
--state		string		Filter by state: running, stopped, pending, etc.
--instance-type		string		Filter by exact instance type
--instance-family		string		Filter by instance family (e.g., m7i)
--tag		strings		Filter by tag key=value (repeatable)
--job-array-id		string		Filter by job array ID
--job-array-name		string		Filter by job array name
--sweep-id		string		Filter by sweep ID
--sweep-name		string		Filter by sweep name

---

spawn status

Show status of an instance or parameter sweep.

spawn status <instance-id-or-name>

Connects to the instance via SSH and runs spored status, reporting lifecycle state, TTL countdown, idle timer, and completion signal.

Examples:

spawn status my-job
spawn status i-0abc123

Flags:

Flag	Type	Default	Description
--check-complete	bool	false	Exit with standardized codes: 0=complete, 1=failed, 2=running, 3=error

--check-complete inspects the instance's completion file (spawn:completion-file, default /tmp/SPAWN_COMPLETE): 0 if present, 1 if it reports a failure status, 2 if absent (still running), 3 on error. Works for a single instance (v0.36.6+) and, via spawn sweep status --check-complete, for parameter sweeps.

---

spawn connect

Open an SSH session to an instance, or run a single command and return.

spawn connect <instance-id-or-name> [-- <command>...]
spawn ssh <instance-id-or-name>

Resolves the instance by ID or name, finds the SSH key automatically from ~/.ssh/, and invokes ssh. Falls back to AWS Session Manager if no public IP is available or --session-manager is set. For DCV app streaming instances, opens the browser session instead.

Pass -- <command> to run a command non-interactively (one-shot mode). The command is wrapped in bash -c '...' on the remote side, so compound operators (&&, ;, &, pipes) and background jobs work correctly. Single quotes in the command are escaped automatically.

If the instance is stopped or hibernated, spawn automatically starts it and waits for it to reach running before connecting. Use --no-start to disable this behaviour.

When multiple instances share the same name, connect prefers the running instance. If multiple running instances share the name, the command fails with a list of the running instances so you can specify an ID.

Examples:

spawn connect my-job                                               # interactive shell
spawn ssh i-0abc123
spawn connect my-job --user ubuntu --port 2222
spawn connect my-job --session-manager

# One-shot mode — single command
spawn connect my-job -- 'tail -20 /tmp/run.log'
spawn connect my-job -- 'aws s3 cp s3://bucket/run.sh /tmp/run.sh'

# Compound commands work — remote shell interprets && ; & etc.
spawn connect my-job -- 'cmd1 && cmd2'
spawn connect my-job -- 'nohup bash /tmp/run.sh > /tmp/run.log 2>&1 &'
spawn connect my-job -- 'aws s3 cp s3://bucket/run.sh /tmp/ && bash /tmp/run.sh'

Flags:

Flag	Type	Default	Description
--user	string	ec2-user	SSH username
--key	string	(auto-detect)	SSH private key path
--port	int	22	SSH port
--session-manager	bool	false	Use AWS Session Manager instead of SSH
--no-start	bool	false	Do not automatically start a stopped/hibernated instance

---

spawn extend

Extend an instance's TTL.

spawn extend <instance-id-or-name> <duration>
spawn extend <duration> --job-array-id <id>
spawn extend <duration> --job-array-name <name>

TTL extension is anchored to the original launch time — extending adds to the current deadline, not to the current wall clock. After updating the tag, spored is signaled to reload its configuration.

Examples:

spawn extend my-job 4h
spawn extend i-0abc123 1d
spawn extend 2h --job-array-name workers

Flags:

Flag	Type	Default	Description
--job-array-id	string		Extend all instances in job array by ID
--job-array-name	string		Extend all instances in job array by name

---

spawn stop / start / hibernate

Stop, start, or hibernate an instance.

spawn stop <instance-id-or-name>
spawn start <instance-id-or-name>
spawn hibernate <instance-id-or-name>

All three commands accept --job-array-id or --job-array-name to operate on an entire job array at once. stop and hibernate preserve EBS volumes — to permanently destroy an instance, use spawn terminate.

---

spawn terminate

Permanently terminate an instance. Unlike stop/hibernate (which preserve EBS volumes), terminate destroys the instance and its EBS volumes. This is irreversible, so it prompts for confirmation by default.

spawn terminate <instance-id-or-name>
spawn terminate --job-array-id <id>
spawn terminate --job-array-name <name>

Examples:

spawn terminate my-instance         # prompts, then terminates (EBS destroyed)
spawn terminate my-instance -y      # skip the confirmation prompt
spawn terminate --job-array-name workers   # terminate the whole array

Flags:

Flag	Short	Type	Default	Description
--yes	-y	bool	false	Skip the confirmation prompt
--job-array-id		string		Terminate all instances in job array by ID
--job-array-name		string		Terminate all instances in job array by name

---

spawn cancel

Cancel a parameter sweep and terminate all associated instances.

spawn cancel --sweep-id <sweep-id>

---

spawn sweep

Manage parameter sweeps after they have been launched.

spawn sweep <subcommand>

Subcommand	Usage	Description
list	spawn sweep list	List parameter sweeps
status	spawn sweep status <sweep-id>	Show sweep progress and instance breakdown
cancel	spawn sweep cancel <sweep-id>	Cancel and terminate all sweep instances
resume	spawn sweep resume <sweep-id>	Resume an interrupted sweep from checkpoint
collect	spawn sweep collect <sweep-id>	Download and aggregate results

spawn sweep list flags:

Flag	Type	Default	Description
--status	string	(all)	Filter by status: RUNNING, COMPLETED, FAILED, CANCELLED
--last	int	20	Show last N sweeps
--since	string		Show sweeps created after date (YYYY-MM-DD)
--region	string		Filter by region

spawn sweep status flags:

Flag	Type	Default	Description
--check-complete	bool	false	Exit with standardized codes: 0=complete, 1=failed, 2=running, 3=error

spawn sweep resume flags:

Flag	Type	Default	Description
--max-concurrent	int	0	Override max concurrent instances (0 = use original)
--detach	bool	false	Run orchestration in Lambda

spawn sweep collect flags:

Flag	Short	Type	Default	Description
--output-file	-f	string	results.json	Output file path
--format		string	json	Output format: json, csv, jsonl
--s3-prefix		string		Custom S3 prefix
--metric		string		Metric field for ranking
--best		int	0	Show top N results (0 = all)
--regions		string		Comma-separated regions to collect from

---

spawn collect-results

Collect and aggregate results from a parameter sweep.

spawn collect-results --sweep-id <sweep-id>

Flags:

Flag	Short	Type	Default	Description
--sweep-id		string	(required)	Sweep ID to collect from
--output	-o	string	results.json	Output file path
--format		string	json	Output format: json, csv, jsonl
--s3-prefix		string		Custom S3 prefix for results
--metric		string		Metric field name for ranking results
--best		int		Show only top N results
--regions		string		Comma-separated regions to collect from

---

spawn schedule

Manage scheduled parameter sweep executions.

spawn schedule <subcommand>

Subcommand	Usage	Description
create	spawn schedule create <params-file>	Schedule a sweep for future execution
list	spawn schedule list	List all schedules
cancel	spawn schedule cancel <schedule-id>	Cancel a schedule
update	spawn schedule update <schedule-id>	Update schedule timing

spawn schedule create flags:

Flag	Type	Default	Description
--at	string		One-time execution time (ISO 8601, e.g., 2026-06-01T14:00:00Z)
--cron	string		Recurring cron expression (e.g., 0 9 * * MON)
--name	string		Schedule name
--max-executions	int		Maximum number of executions
--end-after	string		Stop after duration (e.g., 30d)
--timezone	string		Timezone for cron (e.g., America/New_York)

---

spawn queue

Manage batch job queues for sequential workload execution.

spawn queue <subcommand>

Subcommand	Usage	Description
status	spawn queue status <instance-id>	Show queue execution status
results	spawn queue results <queue-id>	Download results from S3
template list	spawn queue template list	List available queue templates
template show	spawn queue template show <name>	Show template details
template generate	spawn queue template generate	Generate config from template

spawn queue results flags:

Flag	Short	Type	Default	Description
--output	-o	string	.	Output directory for results

---

spawn autoscale

Manage auto-scaling job arrays that grow and shrink in response to queue depth or custom metrics.

spawn autoscale <subcommand>

Subcommand	Usage	Description
launch	spawn autoscale launch	Launch an auto-scaling group
update	spawn autoscale update <name>	Update desired/min/max capacity
status	spawn autoscale status [name]	Show group status (omit name for all)
health	spawn autoscale health <name>	Show instance health details
pause	spawn autoscale pause <name>	Pause reconciliation (keep instances running)
resume	spawn autoscale resume <name>	Resume reconciliation
terminate	spawn autoscale terminate <name>	Terminate group and all instances
set-policy	spawn autoscale set-policy <name>	Set or update queue-depth scaling policy
set-metric-policy	spawn autoscale set-metric-policy <name>	Set metric-based scaling policy
scaling-activity	spawn autoscale scaling-activity <name>	Show recent scaling events
metric-activity	spawn autoscale metric-activity <name>	Show recent metric-based events
add-schedule	spawn autoscale add-schedule <name>	Add a scheduled scaling action
remove-schedule	spawn autoscale remove-schedule <name> <schedule-name>	Remove a scheduled action
list-schedules	spawn autoscale list-schedules <name>	List all scheduled actions

Persistent flags (all subcommands):

Flag	Type	Default	Description
--table	string	spawn-autoscale-groups	DynamoDB table name
--env	string	production	Environment: production or staging

spawn autoscale launch flags:

Flag	Type	Default	Description
--name	string	(required)	Group name
--instance-type	string	(required)	EC2 instance type
--ami	string	(required)	AMI ID
--desired-capacity	int	(required)	Initial desired instance count
--min-capacity	int	0	Minimum instance count
--max-capacity	int	(desired×2)	Maximum instance count
--spot	bool	false	Use Spot instances
--job-array-id	string	(auto)	Job array ID
--key-name	string		SSH key pair name
--subnet-id	string		Subnet ID
--security-groups	strings		Security group IDs
--iam-profile	string		IAM instance profile
--user-data	string		User data script (base64)
--tags	key=value		Additional EC2 tags
--scaling-policy	string		queue-depth or metric
--queue-url	string		SQS queue URL for queue-depth scaling
--target-messages-per-instance	int	10	Scale up when queue exceeds this per instance
--scale-up-cooldown	int	60	Seconds between scale-up actions
--scale-down-cooldown	int	300	Seconds between scale-down actions
--metric-policy	string		Metric name for metric-based scaling
--target-value	float		Target metric value
--metric-name	string		CloudWatch metric name
--metric-namespace	string		CloudWatch metric namespace
--metric-statistic	string	Average	CloudWatch statistic
--metric-period	int	300	Metric evaluation period (seconds)

spawn autoscale update flags: --desired-capacity, --min-capacity, --max-capacity

spawn autoscale set-policy flags: --scaling-policy, --queue-url, --queue (multiple), --queue-weight, --target-messages-per-instance, --scale-up-cooldown, --scale-down-cooldown, --none (remove policy)

spawn autoscale set-metric-policy flags: --metric-policy, --target-value, --metric-name, --metric-namespace, --metric-statistic, --metric-period, --none

spawn autoscale add-schedule flags: --name, --schedule (cron expression), --desired-capacity, --min-capacity, --max-capacity

---

spawn burst

Launch cloud instances to join a local job array (cloud bursting).

spawn burst --job-array-id <id> [flags]

Flags:

Flag	Type	Default	Description
--job-array-id	string	(required)	Job array ID to join
--job-array-name	string		Job array name
--count	int	1	Number of instances to burst
--instance-type	string	t3.micro	Instance type
--spot	bool	false	Use Spot instances
--ami	string		AMI ID
--key-name	string		SSH key pair name
--subnet-id	string		Subnet ID
--security-groups	strings		Security group IDs

---

spawn slurm

Convert and run Slurm batch scripts on EC2.

spawn slurm <subcommand> <script.sbatch>

Translates #SBATCH directives to spawn flags. Supports --nodes, --ntasks, --mem, --gres, --time, --partition, and others.

Subcommand	Description
convert <script>	Convert sbatch script to spawn launch flags
estimate <script>	Estimate cloud cost for the job
submit <script>	Convert and launch on EC2

spawn slurm convert flags:

Flag	Short	Type	Default	Description
--output	-o	string		Output file for converted parameters
--force-yes		bool	false	Skip confirmation prompts

spawn slurm submit flags: inherits --spot

---

spawn stage

Manage data staging for multi-region sweeps.

spawn stage <subcommand>

Subcommand	Usage	Description
upload	spawn stage upload <local-path>	Upload data to regional S3 buckets
list	spawn stage list	List staged datasets
estimate	spawn stage estimate	Estimate transfer cost savings
delete	spawn stage delete <id>	Delete staged data

spawn stage upload flags: --regions, --dest, --sweep-id

---

spawn cost

View cost breakdown for a parameter sweep.

spawn cost breakdown <sweep-id>

Shows per-instance resource costs, utilization, and budget status.

---

spawn dns

Manage DNS records for instances.

spawn dns <subcommand>

Subcommand	Description
register	Register a DNS record for an instance
delete	Delete a DNS record
list	List DNS records

---

spawn notify

Manage Slack and Teams notification registrations for instances.

spawn notify <subcommand>
spawn bot <subcommand>

spawn bot is an alias for spawn notify.

Subcommand	Description
register	Register an instance for chat bot control
deregister	Remove a chat bot registration
enable	Re-enable bot access for an instance
disable	Temporarily disable bot access
list	List bot registrations for a workspace
workspace-add	Register a Slack/Teams workspace's bot token
workspace-remove	Remove a workspace registration
workspace-list	List registered workspaces

spawn notify register flags:

Flag	Type	Description
--platform	string	(required) slack or teams
--user	string	User email address (resolved to platform user ID)
--user-id	string	Platform user ID (alternative to --user)
--workspace-id	string	Slack workspace ID or Teams tenant ID
--instance-id	string	Instance to register
--nickname	string	Display name for the instance in chat
--connect-code	string	Self-registration code from /spore connect
--role-arn	string	Cross-account IAM role ARN

spawn notify workspace-add flags: --platform, --workspace-id, --bot-token, --signing-secret

---

spawn alerts

Manage alerts for sweeps and schedules.

spawn alerts <subcommand>

Subcommand	Usage	Description
create	spawn alerts create <sweep-id>	Create a new alert
list	spawn alerts list	List all alerts
delete	spawn alerts delete <alert-id>	Delete an alert

spawn alerts create flags:

Flag	Type	Default	Description
--on-complete	bool	false	Alert on completion
--on-failure	bool	false	Alert on failure
--cost-threshold	float	0	Alert when spend exceeds this amount
--long-running	int	0	Alert after this many hours
--instance-failed	bool	false	Alert when an instance fails
--email	string		Email destination
--slack	string		Slack webhook URL
--sns	string		SNS topic ARN
--webhook	string		Webhook URL

---

spawn validate

Validate compliance and infrastructure configuration.

spawn validate [flags]

Flags:

Flag	Type	Default	Description
--nist-800-171	string		Validate NIST 800-171 compliance
--nist-800-53	string		Validate NIST 800-53 (low, moderate, high)
--infrastructure	bool	false	Validate infrastructure resources
--instance-id	string		Specific instance to validate
--region	string		AWS region
--output	string	text	Output format: text, json

---

spawn pipeline

Manage multi-stage compute pipelines.

spawn pipeline <subcommand>

Chains stages together: when one stage completes (via --on-complete), the next launches automatically with its results.

See the pipelines guide for full workflow documentation.

---

spawn plugin

Manage plugins for instance customization at launch time.

spawn plugin <subcommand>

Subcommand	Description
install	Install a plugin
list	List installed plugins
status	Show plugin status
remove	Remove a plugin
validate	Lint one or more plugin.yaml files offline (no install): spawn plugin validate <path>...

See the plugins guide for usage.

---

spawn instance-config

Read or write runtime configuration on a running instance via SSH.

spawn instance-config <instance-id-or-name> <action> [key] [value]
spawn config <instance-id-or-name> <action> [key] [value]

spawn config is an alias. Actions: get, set, list.

Examples:

spawn config my-job list
spawn config my-job get idle-timeout
spawn config my-job set idle-timeout 2h

---

spawn team

Manage team-shared instance access.

spawn team <subcommand>

See the teams guide for full workflow documentation.

---

spawn availability

Display historical launch success/failure statistics for an instance type.

spawn availability --instance-type <type> [--regions <regions>]

Flags:

Flag	Type	Description
--instance-type	string	(required) Instance type to check
--regions	string	Comma-separated regions

---

spawn defaults

Manage launch defaults stored in ~/.spawn/config.yaml.

spawn defaults <subcommand>

---

spawn ami

Manage spawn-managed AMIs.

spawn ami <subcommand>

Subcommand	Usage	Description
list	spawn ami list	List spawn-managed AMIs
create	spawn ami create <instance-id-or-name>	Create an AMI from a running instance

spawn ami list flags:

Flag	Type	Description
--region	string	AWS region (default: current region)
--stack	string	Filter by spawn:stack tag
--version	string	Filter by spawn:version tag
--arch	string	Filter by architecture: x86_64, arm64
--gpu	string	Filter by GPU support: true, false
--deprecated	bool	Include deprecated AMIs

spawn ami create flags:

Flag	Type	Description
--name	string	(required) Name for the AMI
--description	string	Description
--tag	strings	Tags in key=value format (repeatable)
--reboot	bool	Reboot instance before creating (default: no-reboot)
--wait	bool	Wait for AMI to become available

---

spawn capacity-block

Purchase EC2 Capacity Blocks for ML from an offering discovered with truffle capacity-blocks.

spawn capacity-block purchase <offering-id> [flags]

⚠️ A Capacity Block is billed up front and is non-refundable — the full block duration is charged at purchase, making this the single most expensive action spawn can take. The purchase requires you to type three confirmations (the exact price, purchase <offering-id>, and an acknowledgement phrase) and refuses to run on a non-interactive terminal — there is no --yes bypass. Use --dry-run first to preview the price and terms without buying anything.

The offering's instance type, count, and duration must be supplied so the exact offering can be re-validated and its current price re-confirmed immediately before purchase. This is step 2 of the Capacity Block flow; launch into the reservation afterward with spawn launch --reservation-id … --capacity-block or lagotto launch --at <block-start>.

Examples:

# Preview only — no charge, no write API call
spawn capacity-block purchase cbo-0abc123 --instance-type p5.48xlarge \
  --count 1 --duration-hours 24 --region us-east-1 --dry-run

# Real purchase (prompts for the three typed confirmations)
spawn capacity-block purchase cbo-0abc123 --instance-type p5.48xlarge \
  --count 1 --duration-hours 24 --region us-east-1

spawn capacity-block purchase flags:

Flag	Type	Default	Description
--instance-type	string		Instance type of the offering (required, e.g. p5.48xlarge)
--duration-hours	int		Capacity Block duration in hours (required)
--count	int	1	Number of instances in the block
--region	string		AWS region of the offering (required)
--platform	string	Linux/UNIX	Instance platform
--dry-run	bool	false	Preview the price and terms without purchasing
--tag	strings		Tag to apply to the reservation (key=value; repeatable)

---

spawn fsx

Manage spawn-managed FSx Lustre filesystems.

spawn fsx <subcommand>

Subcommand	Usage	Description
list	spawn fsx list	List all spawn-managed FSx filesystems
info	spawn fsx info <filesystem-id>	Show filesystem details and cost estimate
delete	spawn fsx delete <filesystem-id>	Delete a filesystem

spawn fsx delete flags:

Flag	Type	Default	Description
--export-first	bool	false	Export data to S3 before deleting
--yes	bool	false	Skip confirmation prompt

---

spawn app

Launch and manage catalog applications via NICE DCV browser streaming.

spawn app <subcommand>

Subcommand	Usage	Description
list	spawn app list	List all streamable applications in the catalog
launch	spawn app launch <app-name>	Launch a catalog application in a browser tab

spawn app list prints each app's name, description, GPU requirement, recommended instance families, and license type.

spawn app launch flags:

Flag	Type	Default	Description
--name	string	<app>-<timestamp>	Session name
--instance-type	string	(catalog default)	Override instance type
--region	string	(AWS config)	AWS region
--spot	bool	false	Use Spot pricing
--ttl	string		Hard termination deadline (e.g. 8h)
--idle-timeout	string	(catalog default)	Stop when DCV has no clients for this duration
--no-open	bool	false	Write session file but do not open browser automatically

After launch, spawn connect <name> reconnects to an existing session — waking the instance if stopped and opening the browser tab.

---

spawn version

Print version, build date, and git commit.

spawn version

---

spawn completion

Generate shell completion scripts.

spawn completion <shell>

Supported shells: bash, zsh, fish, powershell

Setup examples:

# bash
spawn completion bash > /etc/bash_completion.d/spawn

# zsh
spawn completion zsh > "${fpath[1]}/_spawn"

# fish
spawn completion fish | source