Raspberry Pi

WARNING: Raspberry Pi support has been dropped as of version 2.0 due to its underperforming hardware and the critical requirements for the latest Ai models including Whisper and Speecht5. It is possible to install and run Kenzy on a Raspberry Pi but the performance in our tests was undesirable (5 to 7 seconds for voice transcription).

Make sure you start with the basic installation before coming to this page..

Video

Make sure you enable the camera in the Pi if that's what you're using. It's simple to turn on. Use the "raspi-config" command, or if you're on the command line then try out:

sudo -s raspi-config

Audio (Optional)

While this is 100% optiona, I tend to like the pulseaudio setup since it gives a bit more control (in my opinion) over the default setup. The installation is pretty simple. here are the packages and they are totally optional.

sudo apt-get install libatlas-base-dev \
  pulseaudio \
  pamix \
  pavucontrol \
  libpulse-dev

Next, you shouldn't need it, but just in case I'm listing it here for your information in the event you need to modprobe the sound driver for the onboard output on the Pi 4.

sudo modprobe snd_bcm2835 

We may need also to add a reference for a library in our .bashrc file to meet some of the dependencies of the software that is required. This should run under the user that will be executing the Kenzy program. (You can also set this at runtime by prefixing it in the command call if you're not using a standard logged in user to start it up.)

echo "" >> ~/.bashrc
echo "export LD_PRELOAD=/usr/lib/arm-linux-gnueabihf/libatomic.so.1" >> ~/.bashrc

Also, if you followed the general steps then you should have pulse audio installed now. You will want to make sure that it is started before continuing.

pulseaudio --start

---

Help & Support

Help and additional details is available at https://kenzy.ai