Over six years ago I asked a question about existing voice control solutions, which was closed a few days later as opinion-based. At the time I didn't care about it very much, since by then I had already selected a solution for my problem. However, I notice it remains the top result when searching for "voice control", so it may be beneficial to reopen it.
What would it take to fix this question? I would say it is already objective enough — it specifies a concrete problem (voice recognition), target platform (Linux) and a prioritized requirement list (ability to recognize Brazilian Portuguese, failing that English in a foreign accent, and preferably a self-contained local module over cloud-based) — but obviously I'm missing something.