Using Rasa core to only return utterance or action instead of performing it

I want to use rasa core like rasa-nlu, I want it to return the next utterance or action given conversational data instead of printing it out to the channel, so for example i can map a given utterance to a recording and play it when i get to the point.

run rasa core as a server and use /parse and /continue endpoints to receive the next action.

Also in domain.yml add action_factory: remote

+1 for @souvikg10 's comment. Also this will be the standard way to run rasa core as of 0.11