RaspberryPI | Node | DM API

Setup

Create the image and burn it on a SD card

Create a new image of Raspberry PI OS Lite (32-bit) with Raspberry Pi Imager

Set the hostname Note: In this doc, the hostname will be voiceflow

Enable SSH and configure your username/Password and Wifi if you want to use it Note: In this doc, the username will be pi

SSH into the Raspberry PI

Power on the Raspberry PI and on your computer, open a terminal. You want to remotely access your Raspberry over SSH with the credentials you’ve set previously.

ssh pi@voiceflow.local

Note: If you want, you can add your authorized_keys files the .ssh directory on the Raspberry PI to avoid typing your password every time you want to connect to it.

Change password

If you want to change the password previously set:

sudo passwd pi

raspi-config

sudo raspi-config

expand the partition set graphic memory to 16 reboot

Update the system and install Git CLI

sudo apt-get update && sudo apt-get upgrade -y
sudo apt-get install -y git

Install NodeJS

sudo su
curl -fsSL https://deb.nodesource.com/setup_18.x | bash -
sudo apt-get install -y nodejs
su pi

Check the install with:

node --version
npm --version

Clone the repo

git clone https://github.com/voiceflow-gallagan/voiceflow-raspi-node.git

Install RaspiAudio Card drivers (only for RaspiAudio HAT)

wget -O - mic.raspiaudio.com | bash

Reboot the Raspberry PI

Use the following command to check if the card is detected:

wget -O - test.raspiaudio.com | bash

Install PM2 to autostart the app

npm install pm2@latest -g

Config PM2 to start on boot

pm2 startup

Install python3-pip

sudo apt install python3-pip

MPG321 | Speaker

sudo apt-get install mpg321
sudo apt-get install libasound2-dev

Install the app packages

cd voiceflow-raspi-node
npm install

Edit .bashrc

sudo nano .bashrc

Add the following line at the end of the file:

cd voiceflow-raspi-node

Create the .env file at the root of the project

PROCUPINE_ACCESS_KEY = "YOUR_ACCESS_KEY"
VF_API_KEY="VF.DM.XYZ"
VF_VERSION_ID="development"
TUNNEL_SUBDOMAIN="vfassistant"
PORT=3978

PROCUPINE_ACCESS_KEY: Get your access key from Picovoice Console VF_API_KEY: Get your project API key from Voiceflow Creator TUNNEL_SUBDOMAIN: The subdomain you want to use for your tunnel.

Start the app

At the root of the project, run:

npm start

On first start, the PM2 list will be updated. You can check the PM2 list with:

pm2 l

As PM2 start on boot, if you reboot the Raspberry PI the app will autostart and listen for the wakeword.

More info

You can create your own Wakeword model with Picovoice Console and change the path in the wakeword.js file

let keywordPaths = 'wakeword/Hey-Voice-flow_en_raspberry-pi_v2_1_0.ppn' // program["keyword_file_paths"];

You can use/download another Vosk model from here: https://alphacephei.com/vosk/models

Localtunnel is used to create a tunnel to your local machine. You can change the subdomain in the .env file. This is useful if you want to send an event to your local machine from the outside using the /api/messages endpoint. When starting the app, you will see your tunnel url in the console.

/api/messages?event=hello