James McClain has managed to get voice recognition working on GNU/Linux. You can now open sites, ask questions and perform other tasks just by voice. While initially developed for Ubuntu it is distro agnostic and can be used by other distributions as well.
McClain explains that the program has 4 parts which include, 'handling hotkeys, transcribing speech, command recognition and the actual command scripts.
A bash script handles the Hotkeys and the transcribing speech is provided by Google.
He tells me:
Command recognition is handled by a c program that I wrote, I made it to be able to recognize many different ways of saying a command (fuzzy matching), it recognizes commands based on dictionaries that look for example like this.
That allows someone to say open firefox, start firefox, or run firefox, and it will still run the same command.
You can see the app in action in the video below.