Skip to main content
Persistent volumes survive instance termination. Your AI agent can create volumes, check available regions, list existing storage, and clean up unused volumes.

​
”I need a 500GB volume for my training dataset. What regions are available?”

The agent checks regions and creates the volume:
  1. list_storage_regions — returns available regions with capacity
  2. create_volume — creates the volume in your chosen region
Available storage regions: Which region do you want? (I’d suggest us-east-1 if your instances are in the US.)
After you reply “us-east-1”:
  1. create_volume — creates training-data (500 GB) in us-east-1
Created volume training-data (500 GB) in us-east-1. Mount it to any instance in the same region from the dashboard or CLI.

​
”Show me all my volumes and how much storage I’m using.”

The agent lists your storage:
  1. list_volumes — returns all volumes with sizes, regions, and status
You have 3 volumes: Total storage: 800 GB. old-experiment has been idle for 2 weeks and is in a different region from your instances. Want me to delete it?

​
”Delete old-experiment and create a new 1TB volume for my image dataset.”

The agent handles both operations:
  1. delete_volume — deletes old-experiment (unused credits refunded)
  2. list_storage_regions — confirms region availability for 1TB
  3. create_volume — creates image-dataset (1000 GB) in us-east-1
Deleted old-experiment (100 GB) in eu-central-1. Unused time refunded to your balance. Created image-dataset (1,000 GB) in us-east-1. Updated storage summary:
  • training-data: 500 GB (us-east-1)
  • model-checkpoints: 200 GB (us-east-1)
  • image-dataset: 1,000 GB (us-east-1)
Total: 1,700 GB across 3 volumes.

​
Tools used in this workflow