Deployment setup

Smelter can be deployed in various ways, depending on your platform, used features and whether it is used standalone or via Membrane Framework plugin.

Standalone server

Standalone server can be deployed by:

Each of SDKs listed below implements a way to automatically provision a Smelter server. However, standalone guides above will still be useful even if you are using Smelter from TypeScript or Elixir (Membrane Framework). For example, you might want to build binaries or deploy the server yourself, and just use SDK to manage that instance.

TypeScript SDK - Node.js

When using TypeScript SDK, the Node.js process will execute only the React code, but actual rendering and media processing will happen in a separate process. By default, SDK downloads compositor binaries and starts a new Smelter servers on its own. However, you can also connect to an existing instance that you start yourself.

First, you need to decide:

Where will your React code run?
Where will your Smelter server run?

The easiest and the most practical scenario would be to:

For production run both Node.js and Smelter server in the same Docker container.
For development run Node.js on your machine and Smelter server locally (either directly or in a Docker container).

TypeScript SDK - Browser (WASM)

When running Smelter in browser everything is purely client-side, so there is no need to deploy anything, but there are some additional steps you need to take to configure your project.

Membrane Framework plugin

To use Smelter in a Membrane pipeline you can use membrane_live_compositor_plugin that wraps our server. For instructions how to deploy Membrane based applications check out their documentation.

The plugin serves as a bridge between Membrane and the Smelter server, which must still be deployed. By default, the plugin will download the necessary binaries and launch the server locally, but you must ensure that all software and hardware requirements are met.

Refer to standalone guides for more information: