Update: It turned out that nothing I came across fit my needs. So in the end, I ended up writing my own program which utilizes the SAPI APIs in Windows and it works perfectly for what I wanted. Surprisingly, it only took me about an hour to write. For search engine hits to this page; If I ever get it polished enough to release, I will come back and update this post with a link to download the program and possibly the source code as well.


When I used Linux as my only operating system across my devices, I found a Python program called Voximp (abandoned). A current fork of which is Voicely.

This software basically allowed you to create custom phrases or sentences that when spoken would execute either a command or emulate keypresses. It could do this without an internet connection.

The possibilities were endless and it improved productivity substantially, and was easy enough to have work on my different computers.

Now I'm in a mixed OS environment again, and must work primarily on Windows 10. I have yet to find a free replacement that works on Windows.

Voximp and its forks rely on libraries that become extremely heavy weight when brought to Windows, and do not provide full functionality.

Does anybody know of a replacement customizable voice/speech-recognition software which can execute any command I choose, when I say something specific?

  • Must not require an active internet connection to recognize (though relying on the internet to update language models is fine)
  • Open source preferred but not required
  • Premium or free
  • An API for the customization would suffice (C/C++ only though)
  • Lightweight preferred, but not required
  • could you provide some examples of day to day tasks that you handled with these commands ? just curious, seems great software
    – StefanS
    Commented May 14, 2016 at 19:36
  • Some examples of simpler things I'd control via voice, granted its been couple years since I had used voice-recognition, I dont have the original config files any longer. GoogleThis:<googles selected text> CalculateThis:<runs selection through calculation script> Play[insert genre here]:<runs script to play music> StopMusic:<formats all harddrives> and for my tablet or old computer DistractMyKid:<saves all work, opens second xserver, runs paint program, forwards work over ssh to other computer>
    – Hydranix
    Commented May 19, 2016 at 19:29
  • Psst where's your code ;)
    – Menasheh
    Commented Dec 2, 2016 at 4:38
  • 1
    It's not quite done, is a little messy, but I'll post it to github very soon bring a link :)
    – Hydranix
    Commented Dec 2, 2016 at 14:29
  • 1
    Now its up there
    – Hydranix
    Commented Dec 2, 2016 at 14:52

2 Answers 2


You can use Dragon NaturallySpeaking:

  • works on Microsoft Windows
  • nonfree
  • allows to define customize voice commands (you need the professional version for Dragon NaturallySpeaking 13 and below)
  • ~500 MB of RAM, and 2 GB of hard drive space
  • not open source

enter image description here

  • This looks promising, $100 price tag feels so unfair when I've had software which was free and I could edit it to meet my needs (which I did). This may be my only option though unless someone has a better idea. Thank you Franck.
    – Hydranix
    Commented May 19, 2016 at 19:32
  • @Hydranix professional version for Dragon NaturallySpeaking 13 and below used to be > $500 :) Commented May 19, 2016 at 22:45

I wrote my own.

Its rough, but it works.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.