shizzlator / speaking-with-a-webpage

Codelab: Speaking with a Webpage

This is a sample solution for the codelab "Speaking with a Webpage".

Step 0 - Configure Compute Engine

Install Java and Maven:

sudo apt-get update
sudo apt-get install -y maven openjdk-8-jdk

The maven jetty plugin listens for http and https connections on ports 8080 and 8443 by default. Open them up on the Compute Engine firewall:

gcloud compute firewall-rules create dev-ports \
    --allow=tcp:8080,tcp:8443 \
    --source-ranges=0.0.0.0/0

Step 1

Create a webapp on Google Compute Engine that can serve static javascript, a static index.html, and a dynamic controller.

When using getUserInput to access microphone input, browsers require the connection to be https. Configure the app for https - for development purposes, a self-signed certificate suffices.

Generate a self-signed SSL cert for now:

keytool -genkey -alias jetty -keyalg RSA \
    -keystore src/main/resources/jetty.keystore \
    -storepass secret -keypass secret -dname "CN=localhost"

Then run:

cd 01-hello-https/
mvn clean jetty:run

Step 2

Use the WebAudio API on the frontend to get microphone data.

Step 3

Stream the raw audio data from the client to the server using WebSockets.

When on an https webpage, websocket connections are required to be secure wss.

Step 4

Use the Speech API to stream transcriptions of speech to the client.

Disclaimer

This is not an official Google product