By Augusto Goncalves (@augustomaia)
A couple weeks ago I did a testing project with voice recognition and voice feedback for AutoCAD View & Data. After that I started asking myself: can we expand this to create a personal assistant? The idea was to leave it on the house, for instance, and then interact with it, so it should be cheap enough...
That was my weekend fun project: setup Raspberry Pi $35 dollar computer with NodeJS to run a simple voice activated personal assistant! This still my first interaction, but seems to work for some basic tasks...
Of course I'm not starting from zero, I'm using Annyang and Responsive Voice libraries for voice. But the really cool piece was Electron: it allow a JavaScript code to run as a desktop application and it's available for many different platforms.
From zero to hero, follow the steps at this Github repository. The config.js file should be adjusted for different phrases, but new commands need implementation at app.js file.
Enjoy! And have fun!
Hello,
Thanks for your example! I tried it in my raspberry pi 3 but is not working. Annyang don't work, it always shows:
Speech Recognition is repeatedly stopping and starting. See http://is.gd/annyang_restarts for tips.
I'm not using Speech Recognition in any other tab.
In the console in debug mode, but I don't have any other tab with the speech recognition.
Do you have any idea?
Thanks a lot!
Posted by: Salchich_ON | 01/01/2017 at 11:08 AM
I had a similar problem on the RPi when the mic (USB) was not properly configured, can you double check that?
Posted by: Augusto Goncalves | 01/09/2017 at 05:32 AM
Hello,
Thanks for the code. I tried it and its giving me the same error as above. My usb mic is working fine. I think electron.js or any other app using chromium are not allowed to use the speech window speech. Can you please let me know if its there any work around
Posted by: hakimuddin thaker | 06/02/2017 at 02:19 AM
I have also tried on my laptop but no luck.
Posted by: hakimuddin thaker | 06/02/2017 at 02:21 AM