> ## Documentation Index
> Fetch the complete documentation index at: https://docs.streamnative.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Set Up Your Environment

> Prepare StreamNative Cloud accounts and permissions to deploy Orca Agents.

This guide walks you through preparing an environment that can deploy Orca Agents on StreamNative Cloud. You will create a dedicated service account, assign the minimum required permissions, and configure tooling for both the CLI and the StreamNative Cloud Console.

## Prerequisites

* [Install](/tools/cli/snctl/snctl-overview#install-snctl) and [configure](/tools/cli/snctl/snctl-overview#configure-snctl) the `snctl` CLI tool.
* [Log in to StreamNative Cloud Console](https://console.streamnative.cloud/?defaultMethod=signup).
* Create or select a [cluster](/cloud/clusters/manage-clusters/cluster#create-a-cluster), [tenant](/cloud/manage-data-streams/tenant#create-a-tenant), and [namespace](/cloud/manage-data-streams/namespace#create-a-namespace) where agents will run.
* Ensure your user has administrator access for the target cluster and organization.

## Create a service account for agents

1. In StreamNative Cloud Console, open the left navigation and choose **Service Accounts**.
2. Click **Create Service Account**.
3. Provide a name and optional description, then click **Confirm**.

<Note title="Tip">
  Leave **Super Admin** unchecked so the account only receives the scoped permissions you grant later.
</Note>

## Authorize the service account

To make the service account work, you need to make the service account granted with proper permissions (`functions`, `packages`, `produce`, and `consume`).

To grant the service account permissions on the namespace level, follow these steps:

<Tabs>
  <Tab title="Cloud Console">
    1. Navigate to the **Namespace Dashboard** page by [switching to the namespace workspace](/cloud/get-started/cloud-console#switch-a-namespace).

    2. On the **Namespace Dashboard** page, click **Configuration** on the left navigation pane.

    3. On the **Namespace configuration** page, click **ADD ROLE**, and select the service account that you want to authorize.

    4. On the drop-down menu below the service account, select the proper permissions to assign to the service account. There are six permissions in total:

    * `consume`: allow the service account to consume messages.
    * `produce`: allow the service account to publish messages.
    * `functions`: allow the service account to submit and manage functions.
    * `sinks`: allow the service account to create and manage sink connectors.
    * `sources`: allow the service account to create and manage source connectors.
    * `packages`: allow the service account to upload and manage pulsar packages. If you want to submit a customized function/connector, then you will need to upload the function/connector’s JAR/NAR/Python file first, which requires the `packages` permission.

          <img src="https://mintcdn.com/streamnative/leDK9vfVV1vzDfM1/media/authorize-sa.png?fit=max&auto=format&n=leDK9vfVV1vzDfM1&q=85&s=104b2ef6f69dd498ca2552ecb26221cf" alt="Authorize Service Account" width="3218" height="1802" data-path="media/authorize-sa.png" />
  </Tab>
</Tabs>

## Grant access to the service account

To grant the underlying infrastructure with access to the newly created service account's OAuth2 key file, you need to create a service account binding via UI.

Go to the `Service Accounts` tab and choose the service account you want to use for running the connector. Clicking on the right button and there will be a `Edit service account bindings` option.

<img src="https://mintcdn.com/streamnative/pWNa1MQhJLhzkRwZ/media/function-binding-sa-1.png?fit=max&auto=format&n=pWNa1MQhJLhzkRwZ&q=85&s=30c14b1d0ee34f0b718a1e43036d46bf" alt="Binding Service Account step-1" width="1999" height="1071" data-path="media/function-binding-sa-1.png" />

Click the `Edit service account bindings`, choose the desired pool member and confirm.

<img src="https://mintcdn.com/streamnative/y4N9nGGIRiHdtP_Y/media/function-binding-sa-2.png?fit=max&auto=format&n=y4N9nGGIRiHdtP_Y&q=85&s=e50110bcbcc1d7df0c39ecd573d86a60" alt="Binding Service Account step-2" width="3456" height="1980" data-path="media/function-binding-sa-2.png" />

You can also enable the `Enable IAM Role Creation` option to create a separate IAM role for the service account.

Now your connector is ready to use the service account in StreamNative environments.

## (Optional) Create a separate IAM role for the service account

StreamNative's I/O components (Pulsar Functions, Pulsar Connectors, and Kafka Connectors) run as cloud-native workloads on AWS, GCP, and Azure infrastructures.

Use the cloud providers' native IAM (Identity and Access Management) services to control access to infrastructure resources such as S3, GCS, and Azure Blob Storage. This removes the need for password-based authentication for the service accounts that run connectors in StreamNative Cloud.

By default, StreamNative uses a single service account per `PulsarCluster` for all I/O components (Pulsar Functions, Pulsar Connectors, and Kafka Connectors) to access underlying infrastructure resources. This means all I/O Components in the same cluster share one service account and the same permissions.

<Warning>
  Using a single service account for all I/O components means all functions and connectors share the same permissions. If one component is compromised, it could access resources intended for other components.

  So it's better to leave the default service account with no permissions and use it for running IO components that do not require access to external resources.

  And create separate service accounts with the minimum required permissions for each IO component that need to access external resources.
</Warning>

To enhance security and improve isolation, you can create a separate IAM role for each service account used to run connectors. This let you grant only the permissions each service account needs.

Create a separate IAM role for your service account:

<Tabs>
  <Tab title="Snctl">
    <Note title="Note">
      This feature is available in [snctl](/tools/cli/snctl/snctl-overview) v1.3.0 or later.
    </Note>

    1. Get the `PoolMember` name and namespace from the `PulsarCluster`

       ```shell theme={null}
       snctl get pulsarcluster <cluster-id> -o yaml
       ```

       In the output, find the `poolMemberRef` block, which looks like:

       ```yaml theme={null}
       poolMemberRef:
         name: <pool-member-name>
         namespace: <namespace>
       ```

           <Note title="Note">
             Multiple clusters may be located in the same `PoolMember`. You do not need to create separate IAM roles for each cluster within the same `PoolMember`.
           </Note>

    2. Create a new `ServiceAccountBinding` that binds the service account to the `PoolMember`

       ```shell theme={null}
       snctl create serviceaccountbinding <binding-name> \
         --pool-member <pool-member-namespace>/<pool-member-name> \
         --enable-iam-account-creation \
         --service-account-name <service-account-name>
       ```

       **AWS only:** You can specify one or more `AWS Assume Role ARNs` that can be assumed by the IAM role created for the service account (repeat the flag for each ARN):

       ```shell theme={null}
       snctl create serviceaccountbinding <binding-name> \
         --pool-member <pool-member-namespace>/<pool-member-name> \
         --enable-iam-account-creation \
         --service-account-name <service-account-name> \
         --aws-assume-role-arns <arn1> \
         --aws-assume-role-arns <arn2>
       ```

           <Note title="Note">
             The IAM role created by StreamNative will include permissions to call `sts:AssumeRole` on `arn1` and `arn2`. You must still add a trust policy on `arn1` and `arn2` to allow the newly created role to assume them, an example likes below:

             ```json theme={null}
             {
                 "Version": "2012-10-17",
                 "Statement": [
                     {
                         "Effect": "Allow",
                         "Principal": {
                             "AWS": "arn:aws:iam::[aws-account-id]:role/StreamNative/sncloud-role/authorization.streamnative.io/iamaccounts/IamAccount-[org-id]-sab-[binding-name]"
                         },
                         "Action": "sts:AssumeRole"
                     }
                 ]
             }
             ```
           </Note>

    3. (Optional) Update an existing `ServiceAccountBinding` to create the IAM role

       ```shell theme={null}
       snctl edit serviceaccountbinding <binding-name>
       ```

    4. Verify the IAM role was created successfully

       ```shell theme={null}
       snctl get serviceaccountbinding <binding-name> -o yaml
       ```

       Expected output (example):

       ```yaml theme={null}
       apiVersion: cloud.streamnative.io/v1alpha1
       kind: ServiceAccountBinding
       metadata:
         creationTimestamp: "2025-06-25T08:45:35Z"
         finalizers:
         - serviceaccountbinding.finalizers.cloud.streamnative.io
         generation: 1
         name: test-admin
         namespace: o-lftqu
         ownerReferences:
         - apiVersion: cloud.streamnative.io/v1alpha1
           kind: ServiceAccount
           name: admin
           uid: 4ef639aa-6278-4863-9a23-f1da50cea448
         resourceVersion: "54707713"
         uid: 918d40ad-551d-416b-a2ae-d41548d6608e
       spec:
         enableIamRoleCreation: true
         poolMemberRef:
           name: azure-eastus-zephyr
           namespace: streamnative
         serviceAccountName: admin
       status:
         conditions:
         - lastTransitionTime: "2025-06-25T08:45:35Z"
           status: "True"
           type: IAMAccountReady
         - lastTransitionTime: "2025-06-25T08:45:35Z"
           status: "True"
           type: ServiceAccountReady
         - lastTransitionTime: "2025-07-16T13:56:17Z"
           status: "True"
           type: ResourceExists
         - lastTransitionTime: "2025-07-16T13:56:17Z"
           status: "True"
           type: PoolMemberReady
         - lastTransitionTime: "2025-07-16T13:56:17Z"
           reason: AllConditionStatusTrue
           status: "True"
           type: Ready
       ```

       In the output, the `status.conditions` array should include a condition with `type: IAMAccountReady` and `status: "True"`, indicating the IAM role was created successfully.<br /><br />

    5. Use the service account when creating I/O components

       <br /><br />You can now select this service account in Console (or use its API key with the CLI) when creating I/O components (Pulsar Functions, Pulsar Connectors, and Kafka Connectors). The components will inherit the permissions granted to the IAM role created in the previous step.
  </Tab>

  <Tab title="Console">
    You can also create a separate IAM role for your service account via StreamNative Cloud Console. Just enable the `Enable IAM Role Creation` option when creating or editing a service account binding.

    <img src="https://mintcdn.com/streamnative/y4N9nGGIRiHdtP_Y/media/function-binding-sa-2.png?fit=max&auto=format&n=y4N9nGGIRiHdtP_Y&q=85&s=e50110bcbcc1d7df0c39ecd573d86a60" alt="Binding Service Account step-2" width="3456" height="1980" data-path="media/function-binding-sa-2.png" />

    **AWS only:** You can specify one or more `AWS Assume Role ARNs` that can be assumed by the IAM role created for the service account (use one line for each ARN)

    <img src="https://mintcdn.com/streamnative/y4N9nGGIRiHdtP_Y/media/function-binding-sa-3.png?fit=max&auto=format&n=y4N9nGGIRiHdtP_Y&q=85&s=33695bce700a9af551bbe4a21250b431" alt="Binding Service Account step-3" width="1528" height="532" data-path="media/function-binding-sa-3.png" />

    <Note title="Note">
      The IAM role created by StreamNative will include permissions to call `sts:AssumeRole` on `arn1` and `arn2`. You must still add a trust policy on `arn1` and `arn2` to allow the newly created role to assume them, an example likes below:

      ```json theme={null}
      {
          "Version": "2012-10-17",
          "Statement": [
              {
                  "Effect": "Allow",
                  "Principal": {
                      "AWS": "arn:aws:iam::[aws-account-id]:role/StreamNative/sncloud-role/authorization.streamnative.io/iamaccounts/IamAccount-[org-id]-sab-[binding-name]"
                  },
                  "Action": "sts:AssumeRole"
              }
          ]
      }
      ```
    </Note>
  </Tab>
</Tabs>

<Note title="Note">
  In Azure, a Managed Identity `sab-[binding-name]-[org-id]` is created;

  In AWS, an IAM role `role/StreamNative/sncloud-role/authorization.streamnative.io/iamaccounts/IamAccount-[org-id]-sab-[binding-name]` is created;

  In GCP, a service account with display name: `IamAccount/[org-id]/sab-[binding-name]` is created.
</Note>

## Set up client tools

### Use `snctl` to manage agents

StreamNative CLI (`snctl`) provides agent-aware commands that are surfaced under `snctl agents`.

1. Set your organization: `snctl config set --organization <org-name>`.
2. Select a cluster context: `snctl context use` and choose the target cluster.
3. Work with agents by using the following commands (pass `--tenant` and `--namespace` as needed):
   * `snctl agents list` – list agents in the current namespace.
   * `snctl agents create` – deploy a new agent from a declarative spec and package.
   * `snctl agents update` – roll out spec or code changes.
   * `snctl agents status` – view runtime status.
   * `snctl agents start|stop|restart` – control the running agent.
   * `snctl agents delete` – remove an agent.
   * `snctl agents get` – show the stored agent spec.
   * `snctl agents trigger` – deliver an ad-hoc test event when supported by the agent implementation.

### Use StreamNative Cloud Console

The StreamNative Cloud Console offers a guided experience for connecting client applications and reviewing agent activity.

Navigate to **Agents** from the left panel to review existing deployments, logs, and status.

## What's next?

* Review the [Orca Agents overview](/agent-engine/agents-overview) to understand runtime architecture and capabilities.
* Package your first agent and deploy it with `snctl` or the Cloud Console.