I would love to see a section in the Readme on how to package a model or be able to use a recipe to package a model.
It took me a while to find out how to package a model. Even though it is a part of the first steps tutorial. After training a model for a while I didn't feel the need to look back at the first steps tutorial. Googling "prodigy package model" or looking in the Readme didn't help either. I found that you have to use the spacy api to package a model in one of Matthews Videos: https://spacy.io/api/cli#package
Sorry if this was confusing! But that's correct – Prodigy itself only handles the annotation workflows and the default recipes use spaCy under the hood. So what you get out at the end if you're using the built-in workflows is just regular spaCy model. Nothing about the exported model data is specific to Prodigy. And you can always plug in your own implementations via custom recipes and use Prodigy without spaCy.
Btw, the spaCy docs also have a more extensive usage guide on saving and loading models, more elegant packaging with entry points etc: https://spacy.io/usage/saving-loading