Use whisper for local speech-to-text. Start with:
whisper
whisper --help
Prefer small files first; keep outputs in the current working directory.