Skip to main content

​
Can I use the AWS CLI as if I were talking to AWS?

Yes—for operations supported by the service adapter. Point the AWS CLI at your Cloud Adapter endpoint. It sends the same AWS service API request and receives an AWS-shaped response; Cloud Adapter calls the selected target service. You do not replace the AWS CLI with a Google Cloud or Azure CLI. For example, an S3 PutObject command can write an object to Google Cloud Storage, while a DynamoDB GetItem command can read an item backed by provisioned Azure Cosmos DB:
These commands assume the adapter, origin-side credentials, mapped bucket and table, and target identity are already configured. AWS_ENDPOINT_URL_S3 and AWS_ENDPOINT_URL_DYNAMODB are the AWS CLI’s service-specific alternatives to repeating --endpoint-url. The overview explains the request path; the AWS CLI guide walks through a safe test. Check the S3 and DynamoDB profiles for exact coverage.

​
Does every cloud work with every other cloud?

No. The cloud logos illustrate the product model. The service reference under each cloud in the sidebar lists declared mappings. Availability and operation coverage must be checked for the specific direction you need.

​
Do I need a Tensor9 API key?

Yes. Cloud Adapter requires a Tensor9 customer account. Use T9_API_KEY when the operator acquires adapters and licenses, renews a license, or uses online customer services such as Explain. Applications do not send this key to the adapter endpoint, and deployed adapters do not retain it.

​
Do I need to replace my cloud SDK?

No. An API-compatible adapter serves the origin SDK’s API, so you configure that client’s endpoint and origin-side authentication rather than replacing the SDK with a target-cloud SDK. For example, an S3 client can use Google Cloud Storage and a DynamoDB client can use Azure Cosmos DB through their respective service adapters. Verify the operations your application uses; some capabilities differ or are unavailable on a backend.

​
Does Tensor9 store my application data?

The selected backend stores the data for a mapping, such as objects in Cloud Storage. Adapters can also need operational state. Review the service mapping and installation design to identify every store that requires backup, access control and lifecycle management.

​
Is Cloud Adapter the same product as Tensor9 BYOC?

They share adaptation technology but have different workflows. Cloud Adapter focuses on using origin service APIs with selected backends. Tensor9 BYOC focuses on delivering your software into another organization’s infrastructure.

​
Can I change backends after storing data?

Do not assume so. A backend change can require a data and identity migration. The upgrade workflow rejects changes without a supported migration. See treating configuration changes as transitions.

​
Does an emulator prove production compatibility?

It helps test the behavior it implements. It does not establish your production IAM configuration, networking, quotas, latency or failure recovery. Use local testing together with a bounded test against the real target service.

​
Can an agent use the same debugging tools?

Yes. Use your existing coding agent with the Tensor9 CLI; no separate agent-specific service is required. Read the agent debugging guide for a safe investigation prompt and explain workflow.

​
Where does the adapter run?

It runs in the deployment scope you choose. The target cloud supplies the networking, process supervision and backend primitives described by that deployment. Cloud Adapter has no resident Tensor9 management control plane. Acquisition and license renewal authenticate to Tensor9, while serving an entitled application request does not require a per-request Tensor9 call.

​
Who can see my data and credentials?

The selected backend stores the data for the mapping. Runtime backend credentials belong to the adapter configuration; origin credentials authenticate application requests. Keep both scoped, rotate them independently and do not place secret values in configuration reviews or agent transcripts.

​
Can I use existing target resources?

Only when the selected mapping and configuration workflow support adopting them. A resource identifier is not proof that data, encryption keys, policies or lifecycle behavior are compatible. Verify ownership and migration requirements before pointing an adapter at populated storage.

​
How does scaling work?

Each Cloud Adapter installation runs one adapter process. Size that process, its network path and its backend together: give the process the CPU and memory its workload needs, and size the backend for the same load. A healthy process does not prove that the backend can handle the workload. Establish a baseline and test failure recovery before increasing load.