I was thinking about creating a kind of Jarvis computer (Iron man). Only then a (little) bit more minimalistic. What I want to built is more of a question/answer type of application (something like Siri) and then extend it to do more stuff like turning lights off and on/ controlling my tv among others. But before I start with this I have a couple of questions regarding the voice control on the Raspberry Pi.

  • Can the Raspberry cpu/gpu handle the workload in order to translate voice to commands
  • If not, is there some sort of web-service I can access with the Raspberry to do the work for me? Perhaps I could buy a cheap Android device and make that do the translation for me (any tutorials??), or some kind of Google web-service?
  • If yes, what is the best way to achieve this? Some Java library, or something else?
  • Can the Raspberry handle multiple wireless microphone inputs (with a powered usb hub of course)?
  • Can the Raspberry handle multiple wireless audio outputs (for answering my questions)?
  • Is there perhaps a compete (or partial) tutorial on this topic? (That would be awsome!)

2 Answers 2


Check out this project that involved speech recognition and a cheap robotic arm: http://www.aonsquared.co.uk/raspi_voice_control

Edit, October 2017: The above URL appears to have gone stale. An archive of the page can be found at https://web.archive.org/web/20130408071125/http://www.aonsquared.co.uk/raspi_voice_control

  • Link answers are taboo here. I know I'm being a hypocrite, but please see this meta post for more information. Commented Apr 23, 2013 at 2:46
  • Yea just summarize the external source and reference the site and a link to it. It is mostly about SEO and nothing else.....
    – Piotr Kula
    Commented Apr 23, 2013 at 10:59
  • The site in question is not mine. Would you all prefer that I merely skipped past the page and not offered any help at all? Commented Apr 23, 2013 at 17:00
  • 1
    Thanks a lot Carl! I think this will get me started and it answers questions 1,3 and 5 Commented Apr 24, 2013 at 8:34
  • page not found :-( Commented Oct 24, 2017 at 16:44

The aonsquared project is actually a bit outdated. Better speech recognition performance can be found using Google's speech recognition or Siri. Someone posted something like this recently on youtube (looks like it takes whatever you say, compares it in a config file, and runs whatever command you tell it to) http://www.youtube.com/watch?v=tQi9dWVUJow

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.