You might want to look at Matrjoschka of phoenix communication | Benjamin Milde. It should make the distinction between the tools more clear. Generally the question will be why you need channels – not just PubSub – when you already have LV running (on top of channels).

Another case for using Phoenix PubSub in a cluster (with src demo) for you Using Phoenix.PubSub as a message-bus for Elixir cluster

I guess you are quite new to elixir and phoenix. So I will try my best - I am not good at explaining

The LiveView model allows live communication between the server and the client browser. But each user is served its own instance of the LiveView. So they do not know of each other.

The PubSub part allows you to send the same message to all subscribers, so you can notify them about changes.

In your case, you would subscribe to a specific PubSub message from the LiveView. And in the LiveView you handle the PubSub message.

You could, for example, broadcast a {:content_updated, 134} message. And the LiveView then knows, that it has to reload the data for that id.

Not full example, but perhaps to explain the difference between the concepts.

Namxam has it right. The only thing I’d add to answer your original question is: if you’re using LiveView you can ignore the Channel documentation. Just focus on LiveView and PubSub. Channels are useful when you’re not using LiveView, e.g. you have a separate frontend.

And that’s because Phoenix.LiveView.Channel builds on top of Phoenix.Channel which is why you might not need to interact with Channels directly when you are already using LiveView

Sometime ago, I published a blog post about how I built a turn-based battle game prototype, which uses PubSub to push state to multiple clients/players. If you are into gaming, the theme might ease you into my poor didactics .

You can check the post here: How I created a JRPG battler with Elixir - the code is also open source if you want to follow along.

This is my go to reference on understanding pubsub for the first time!