gpu registry-credentials

Manage private-registry logins for custom-image launches

Synopsis

Store a private container-registry login once, then launch private images with gpu instances create --image ... --registry-credential . Passwords are encrypted at rest and never returned by any read — the add call is the only time the secret crosses the wire. See the Custom images guide.

gpu registry-credentials add

GPUAI_REGISTRY_PASSWORD=ghp_xxx gpu registry-credentials add \
  --name ghcr-acme --registry ghcr.io --username acme-bot

The password/token is read from the GPUAI_REGISTRY_PASSWORD environment variable — deliberately not a flag, so it can't leak via shell history or the process list. --name is an org-unique handle (duplicates are a 409); --registry is the registry host the login belongs to and must match the image's registry at launch time.

      --name string       Org-unique handle for this login (e.g. ghcr-acme) — required
      --registry string   Registry host the login belongs to (e.g. ghcr.io, docker.io) — required
      --username string   Registry username — required

gpu registry-credentials list

Lists stored logins — id, name, registry, username, created. Passwords are never shown (or returned by the API).

gpu registry-credentials remove

gpu registry-credentials remove <credential-id>

Deletes a stored login. Instances already running keep working; the next launch referencing the deleted id fails with a 422.

Options inherited from parent commands

      --api-base string   API base URL (override with GPUAI_API_BASE env) (default "https://api.gpu.ai/v1")
      --debug             Enable debug logging to stderr
  -o, --output string     Output format: table|json (default table on TTY, json otherwise)

SEE ALSO

• gpu - GPU.ai command-line interface
• gpu instances create - Create a new GPU instance