Remote Instances

A Snowman instance can be hosted to allow remote access. This can be useful in case you want to collaborate with coworkers or provide an instance for a client to connect to. We refer to such instances as remote instances.

Warning

Be advised: We have not yet implemented any security features or authorization. That means every process with access to the host is able to access the API. Progress is tracked in issue #107. We therefore recommend to all users to set up authentication and encryption for their remote instances by themselfs. This can be achieved through reverse proxies like Caddy or nginx.

How it Works

Snowman is a web server and web app packaged with Electron. That means accessing a remote instance is as easy as visiting the website of the remote instance (this can be done with the Snowman client or a normal web browser). By using this setup we avoid version conflicts as a client will always use the frontend provided by the remote instance.

Client Perspective

As a client, you will receive the URL (=address) of the Snowman instance you should connect to. When opening Snowman app, make sure to not start a local instance but selecting REMOTE INSTANCE and entering the provided address.

Provider Perspective

As a provider, you have two options for hosting a Snowman instance.

Headless Mode

Snowman offers a headless mode (does not open a GUI window). Access it by including the --headless flag when starting snowman from the command line. For further command line arguments and how to configure host and port see here.

Example:
./snowman --headless --hostname 192.168.12.34 --port 8123 --storageDirectory /home/user/snowman
In this case, the address http://192.168.12.34:8123 can be used by clients to connect to this instance.

Docker

As an alternative, we provide a fully featured Docker image that runs a Snowman server.

This feature is WIP - progress is tracked in issue #216!