sentimental

Sentimental

This example application demonstrates the power of Media Streams by forking audio from a call, transcribing it, and performing real-time sentiment analysis.

Setup

Create a Google Application that has Google's Speech to Text API and the Cloud Natural Language API enabled.

Create credentials for a Service Account and save them as google_creds.json in the folder next to this README.md file.

Install and configure the server

npm install
npx configure

To run locally you'll need to open a port on your machine, so that Twilio can communicate. You can do this using ngrok.

ngrok http 3000

Run cp .env.example .env to create a .env file with the needed environment variables.

Update your .env file to contain your ngrok URL and the other needed variables.

Note: NODE_ENV is set to production by default.

Set your incoming number to point to your ngrok url (or use the Twilio Console to update your incoming number)

twilio phone-numbers:update +15558765 --voice-url=https://your-ngrok-url.ngrok.io/twiml