Skip to main content

Documentation Index

Fetch the complete documentation index at: https://runcrate.ai/docs/llms.txt

Use this file to discover all available pages before exploring further.

Environments let you organize resources within a workspace. Your AI agent can create, list, and delete environments, and show you what’s running in each one.

​
”Set up staging and production environments for our ML pipeline.”

The agent creates both environments:
  1. list_environments — checks what already exists
  2. create_environment — creates staging
  3. create_environment — creates production
You already have the default main environment. Created two new environments:
  • staging — for testing model updates before promotion
  • production — for live inference workloads
Your workspace now has 3 environments: main, staging, production. Deploy instances to a specific environment from the dashboard or specify it when creating resources.

​
”What resources are running in each environment?”

The agent maps instances and storage to environments:
  1. list_environments — lists all environments
  2. list_instances — lists all instances with their environment assignments
  3. list_volumes — lists all volumes with their environment assignments
main (default)
  • dev-box (RTX 4090) — running, $0.35/hr
  • scratch-volume (100 GB)
staging
  • staging-inference (A100) — running, $1.60/hr
  • staging-models (200 GB)
production
  • prod-inference-1 (H100) — running, $2.50/hr
  • prod-inference-2 (H100) — running, $2.50/hr
  • model-weights (500 GB)
Total burn rate: $6.95/hr Production is 72% of your spend.

​
”Delete the staging environment — we’re done testing.”

The agent checks for active resources first:
  1. list_instances — checks for instances in the staging environment
  2. ssh_execute — optionally checks if the instance is mid-task
staging has 1 running instance (staging-inference) and 1 volume (staging-models). Environments can’t be deleted while they have active resources. Want me to:
  1. Terminate the instance and delete the volume, then delete the environment
  2. Just terminate the instance and keep the volume
After you choose option 1:
  1. delete_instance — terminates staging-inference
  2. delete_volume — deletes staging-models
  3. delete_environment — deletes the staging environment
Terminated staging-inference, deleted staging-models (200 GB), and removed the staging environment. Your workspace now has 2 environments: main, production. Saving $1.60/hr on the instance.
The default main environment cannot be deleted. This is enforced at the database level.

​
Tools used in this workflow

ToolPurpose
list_environments / create_environment / delete_environmentManage environments
list_instances / list_volumesCheck resources per environment
delete_instance / delete_volumeClean up before environment deletion