How to develop a speech recognizer?

Question

Python

I need something that is not standard, not just an api or a service on the Internet.
I need the technology itself.
Let's say I write down my phrases, with my voice, and make associations with text or commands.
How to make it so that not only my voice is recognized, but also the voices of other people with different intonations.
A long time ago there was a phone with a black and white screen, he knew how to learn and execute commands.
Here you need the same thing to make Jarvin :)
What to read? look?

Answer 1

https://habrahabr.ru/post/226143
offhand found such sources.
the task, by the way, is difficult and for a beginner probably unbearable

Answer 2

Technology can be. First, breakdown by time and spectral analysis. I advise you to use wavelets. Then the selection of phonemes (make a dictionary). Make words out of phonemes. But it's a long way. The best thing is just the API (eg Speech API from Microsoft).