NAME
    gcloud alpha ai endpoints - manage Vertex AI endpoints

SYNOPSIS
    gcloud alpha ai endpoints COMMAND [GCLOUD_WIDE_FLAG ...]

DESCRIPTION
    (ALPHA) An endpoint contains one or more deployed models, all of which must
    have the same interface but may come from different models. An endpoint is
    to obtain online prediction and explanation from one of its deployed
    models.

    When you communicate with Vertex AI services, you identify a specific
    endpoint that is deployed in the cloud using a combination of the current
    project, the region, and the endpoint.

GCLOUD WIDE FLAGS
    These flags are available to all commands: --help.

    Run $ gcloud help for details.

COMMANDS
    COMMAND is one of the following:

     create
        (ALPHA) Create a new Vertex AI endpoint.

     delete
        (ALPHA) Delete an existing Vertex AI endpoint.

     deploy-model
        (ALPHA) Deploy a model to an existing Vertex AI endpoint.

     describe
        (ALPHA) Describe an existing Vertex AI endpoint.

     direct-predict
        (ALPHA) Run Vertex AI online direct prediction.

     direct-raw-predict
        (ALPHA) Run Vertex AI online direct raw prediction.

     explain
        (ALPHA) Request an online explanation from an Vertex AI endpoint.

     list
        (ALPHA) List existing Vertex AI endpoints.

     predict
        (ALPHA) Run Vertex AI online prediction.

     raw-predict
        (ALPHA) Run Vertex AI online raw prediction.

     stream-direct-predict
        (ALPHA) Run Vertex AI online stream direct prediction.

     stream-direct-raw-predict
        (ALPHA) Run Vertex AI online stream direct raw prediction.

     stream-raw-predict
        (ALPHA) Run Vertex AI online stream raw prediction.

     undeploy-model
        (ALPHA) Undeploy a model from an existing Vertex AI endpoint.

     update
        (ALPHA) Update an existing Vertex AI endpoint.

NOTES
    This command is currently in alpha and might change without notice. If this
    command fails with API permission errors despite specifying the correct
    project, you might be trying to access an API with an invitation-only early
    access allowlist. These variants are also available:

        $ gcloud ai endpoints

        $ gcloud beta ai endpoints

