There is pretty good write up https://towardsdatascience.com/monkeying-around-with-alexa-routines-3f3f95795922 which explains how someone used a Raspberry Pi to invoke Alexa Routines with Python code by using a Skill called Voice Monkey.
The skill allows you to run Alexa Routines by creating Virtual Triggers/Buttons, which can be invoked by GET requests. Not only can it to run Alexa Routines but the skill also has Text-to-speech capabilities which this forum might find interesting. For me, the more interesting capability, is the the skills ability to allow Alexa to ask Yes or No question, and respond differently based on your answer.
There is nominal yearly fee to get "Pro" features which include Text-to-Speech announcements, sounds, etc.. https://voicemonkey.io/
Using the API to trigger Alexa Routines and Text-to-speech works perfectly in my testing, although I am having some trouble with the Yes/No prompting. The prompt works fine if I respond, but not so much if I do not.
I was a bit surprised to seemingly not see this Amazon skill mentioned anywhere in the forums, given this is SO MUCH easier than how I got Alexa Text-to-speech working on my system.
Is anyone else using the Voice Monkey skill and more importantly, has anyone gotten the Yes/No prompting to work?