> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tensor9.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Single binary

> Run one service adapter process with a simple, explicit operating boundary.

<Frame>
  <img className="block dark:hidden" src="https://mintcdn.com/tensor9/3qwgIru3kCfHriJC/images/diagrams/cloud-adapter-single-binary-light.svg?fit=max&auto=format&n=3qwgIru3kCfHriJC&q=85&s=8a3c07b3183a32a72f74469ad9f9297c" alt="A single Cloud Adapter process receives origin-cloud API calls and translates them to native target services." width="1200" height="600" data-path="images/diagrams/cloud-adapter-single-binary-light.svg" />

  <img className="hidden dark:block" src="https://mintcdn.com/tensor9/3qwgIru3kCfHriJC/images/diagrams/cloud-adapter-single-binary-dark.svg?fit=max&auto=format&n=3qwgIru3kCfHriJC&q=85&s=ab471b1b111e58703c996dc44d1a08c2" alt="A single Cloud Adapter process receives origin-cloud API calls and translates them to native target services." width="1200" height="600" data-path="images/diagrams/cloud-adapter-single-binary-dark.svg" />
</Frame>

The **single binary** deployment runs one service adapter process. It is the smallest useful operating shape: one endpoint, one process to supervise and one configuration to manage.

## When to use it

Use a single binary for local development, evaluation, low-volume workloads or a service where process-level redundancy is provided by the surrounding platform. “Single” describes the adapter process, not the durability of your data.

## What you operate

* The adapter binary and its runtime configuration.
* The endpoint and network path used by callers.
* Runtime credentials and access to destination services.
* Backend data, backups and recovery procedures.

For production, place the process under a supervisor and exercise restart and backend-failure recovery before sending traffic.

## Move to a cluster

When one process is no longer enough for availability or throughput, keep the same service configuration and move to [Cluster](./cluster). Validate shared state, connection limits and rollout behavior before switching traffic.
