Skip to content

How-to: browse the public template catalog and check GPU availability per datacentre

This is a worked task that spans a few commands. The reference pages (rp doc catalog list, rp doc template search, rp doc stock dc, rp doc stock gpu) document the individual flags; this page shows them composed into one workflow.

Goal: list the public community templates you can clone into a pod or serverless endpoint, and find out which datacentres currently have GPUs in stock so you can pin --dc (or choose a region for a cluster) with confidence.

Steps

  1. List the public catalog. --limit, --cursor, --jq, and --json are all valid:

$ rp catalog list --limit 50

Each row carries id, name, image, serverless, and public. To pull just the names:

$ rp catalog list --json --jq '.[].name'

To keep only the serverless-ready entries:

$ rp catalog list --json --jq 'map(select(.serverless))'

To filter your own templates by name substring instead, use rp template search <name-substring> (case-insensitive) — that matches against your account's templates, not the public catalog:

$ rp template search vllm

  1. See which datacentres have GPUs in stock. rp stock dc returns a table with the DATACENTER id, its REGION, GPUS (how many GPU types have any stock there, not how many cards), and the S3_API, GLOBAL_NETWORK, NETWORK_VOLUME_TYPES, and COMPLIANCE columns:

$ rp stock dc

The DATACENTER value is exactly what rp pod create --dc, rp volume create --dc, and rp cluster create --dc take.

  1. Get per-datacentre GPU detail. rp stock dc --json prints the datacentre records including the gpuAvailability array, which lists each GPU type and its per-datacentre availability for every location:

$ rp stock dc --json

This is the source of truth for "which GPU types are in stock where" — the gpuAvailability field already carries the datacentre dimension, so no second query is needed.

Notes

  • rp template and rp catalog are different resources. rp template lists yours — the private/serverless container configs saved under your account. rp catalog list reads the public community catalog, which is list-only (no create/get/delete of its own). Copy an id from the catalog into rp pod create --template-id or rp serverless create --template-id.
  • Use rp stock dc --json and read gpuAvailability for per-datacentre stock. rp stock gpu --product SERVERLESS --json is the wrong tool for that: it reports one availability category per GPU type, not a datacentre breakdown.

  • rp stock gpu is still the right tool when you want the list of in-stock GPU display names (the ID column) to pass to rp pod create --gpu or rp serverless create --gpu. It also supports column-level filters that apply to both the table and --json: --cloud SECURE|COMMUNITY, --vram-gb N (or --vram N, a minimum VRAM), --stock <level>, --cuda <ver>, and --sort <column>.

  • S3_API is shown only in the table, not in --json.