Hmm, yeah, I think if you want to integrate Phoenix into an existing Erlang application, you’ll probably want to use Elixir as the “main” language and use existing Erlang code from the Elixir app. Or if you really need to run Phoenix inside Erlang, you could make a clean cut and write a portion of the app in Elixir, with an easy way to call it (e.g. start a server) from Erlang.
Neither option allows you to use Phoenix without writing Elixir code, though. But since you’re integrating an Elixir lib, writing at least some glue code doesn’t seem like a big deal?
(I’ve got to say that I’m not proficient at all neither in Erlang nor in Elixir – I’ve played around with both, but nothing serious.)