Machine Catalogs

Varity TeamCore Contributors

Varity Cloud exposes three machine catalog scopes. Capacity, regions, configurations, and GPU prices can change independently of this page, so the authenticated catalog and a fresh quote are the only current sources.

The Three Catalogs

Catalog	workload	execution_class	Create with	Access	Monthly maximum
CPU VMs	virtual_machine	cpu_virtual_machine	POST /api/machines	SSH	When present in the quote
GPU VMs	gpu	virtual_machine	POST /api/machines	SSH	No
GPU bare metal	gpu	bare_metal	POST /api/machines	SSH	No

Both query parameters are required. Treat availability as a snapshot, not a reservation.

You may not need a machine

Applications are a managed Varity Cloud workload: you provide a supported repo or image and Varity returns an application URL. Choose a machine only when you need host-level access or an exact accelerator configuration.

Reading a Catalog

curl "https://varity.app/api/deployment-profiles?workload=virtual_machine&execution_class=cpu_virtual_machine" \
  -H "Authorization: Bearer $VARITY_API_KEY"

curl "https://varity.app/api/deployment-profiles?workload=gpu&execution_class=virtual_machine" \
  -H "Authorization: Bearer $VARITY_API_KEY"

Use execution_class=bare_metal for the other GPU scope. Read the selected profile’s identifier, resources, region, images, availability, and price fields from the response you receive. Do not copy a profile id or GPU rate from documentation into automation.

Quote Before Create

• CPU VMs and GPU machines use POST /api/pricing/machine-quote.
• The quote binds the current price and configuration and carries a valid_until deadline.
• If the quote expires or any workload input changes, request another quote.
• A catalog price is discovery data. The accepted quote is the price contract for the create request.

GPU pricing fails closed

This documentation deliberately publishes no GPU rate, cheapest-offer claim, or continuous-use total. If a fresh workload-bound quote is unavailable, no GPU price is available to publish or spend against.

Next

CPU Virtual MachinesCPU VM selection, quote, create, access, and lifecycle.

GPU WorkloadsGPU VMs and bare metal through fresh catalogs and quotes.

API ReferenceAuthentication and the public resource interface.