How to create speech commands data set

Question

I am planning to create a speech recognition network that recognize few words (voice commands) and came across Speech Commands dataset from google. Apart from available dataset I am planning to add few more words like "move", "save" etc, which are not part of the google's dataset.

Is there a way or any other source where i can create/find the dataset for the words ?

As many of the online dataset are available for sentences and speech transcripts i am thinking of writing a scripts that can go through the available transcripts and find the location of the desired word and physically cropping the audio and then padding it to make one second audio file. But this seems very very arduous process and few words like the name of my robot will not be available online.

Can you guys suggest a way ?