gpu secrets
Store secrets once and hand them to instances, functions, and jobs
Store secrets once and hand them to instances, functions, and jobs
Synopsis
Organization secrets are values such as API keys and tokens that you store once and name on a launch. They reach the machine as environment variables of the same name.
Values are encrypted at rest and are never shown again after they are saved. Every read returns names and metadata only. Every create, update, delete, and use is recorded in your organization's audit log.
Use them with gpu instances create --secret NAME, gpu jobs submit --secret NAME, or @function(secrets=["NAME"]) in Python. Coming from Infisical? Run infisical export --format=json > secrets.json and then gpu secrets import --file secrets.json.
Secrets are off by default. On a deployment that does not offer them, these commands fail with this server has no such endpoint. See the secrets guide for how values are encrypted and delivered.
Examples
# Store a token (prompts without echo)
gpu secrets set HF_TOKEN
# Store one from a script
printf %s "$TOKEN" | gpu secrets set WANDB_API_KEY
# Use it
gpu instances create -t a100_80gb --secret HF_TOKEN
Options
-h, --help help for secrets
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 secrets delete - Delete a secret
- gpu secrets get - Show a secret's metadata (the value is never shown)
- gpu secrets import - Import secrets from a .env file or an Infisical export
- gpu secrets list - List secrets (names and metadata, never values)
- gpu secrets set - Create a secret or replace its value