🎤 AI (Speech to Text) Demo

Node.js sample applications that shows IBM Watson Speech to Text service with RingCentral Data (Meeting Recording/Call Recording).

Introduction

This is a Node.JS based application that uses 2 main technologies - IBM Watson's Speech to Text AI Service and RingCentral's Platform to flag certain keywords when spotted in the recording, in this case for complaince reason but you can change that to anything. First we download Call and Meeting recording and detailed teps to achive the same as documented below. Then, we confifure in the GUI - Speech to Text Model for any given languages, 'Complaince Keyworkds' that you want to look for in the trasncribed speech etc and run the app by either direct speech input or by using the audio recordings.

Prerequisites

Configure RingCentral

1. Create an application on RingCentral Developer Portal (sign up for free if you are a new developer)

• Login or create an account if you have not done so already.
• Go to Console/Apps and click 'Create App' button.
• Select "REST API App (most common)" under "What type of app are you creating?". The current project calls the API from Node.JS backend so method fits best.
• Provide App Name, Primary Contact Name
• For "Do you intend to promote this app in the RingCentral App Gallery? (for internal-use only)", choose 'No' if you're creating just for learning purpose.
• Auth : Choose 'Password-based auth flow' for this demo app. For your 'real world' production app, password flow is not recommended. Please find more information here
• 'Issue Refresh token' : Choose 'Yes'
• Then select the following Application permissions: Meetings, Read Call Log, Read Call Recording
• Click 'Create' and your application will be created

2. Clone this GitHub repository and install dependencies including RingCentral Node.JS SDK etc

git clone https://github.com/suyashjoshi/ringcentral-ai-demo
cd <clone repository folder>
npm install

Configure RingCentral Video (rcv.js) API

• You will need to use RingCentral Video API, currently that is in 'early access beta'. You can request access here
• Once you have access, login to App Console and paste the credentials in rcv.js file
• We will call the "Meeting History API" to get the meeting logs and media content URI
• Lastly, we will see the URL to download the .mp4 file in the console. Use web-browser to curl to download the .mp4 file. You then need to convert it to .mp3 or .wav format in order to use this this AI model as it only works with audio file.
• Open the project and navigate to 'rcv.js' file. Here you need to read the comments and update the fields with your credentials that you acquired in the previous step.

npm install
node rcv.js

Note: If you're using RingCentral Meeting API, refer to this page https://developers.ringcentral.com/api-reference/Call-Recordings/readCallRecording

Configure RingCentral Call Log & Recording API (rc-call-recordings.js)

• Open the project and navigate to 'rc-call-recordings.js' file. Here you need to read the comments and update the fields with your credentials that you acquired in the previous step.

npm install
node rc-call-recordings.js

Configure IBM

1. You need an IBM Cloud account.

2. Download the IBM Cloud CLI.

3. Create an instance of the Speech to Text service and get your credentials:

   • Go to the Speech to Text page in the IBM Cloud Catalog.
   • Log in to your IBM Cloud account.
   • Click Create.
   • Click Show to view the service credentials.
   • Copy the apikey value.
   • Copy the url value.

4. In the application folder, copy the .env.example file and create a file called .env

   cp .env.example .env
   

5. Open the .env file and add the service credentials that you obtained in the previous step.

   Example .env file that configures the apikey and url for a Speech to Text service instance hosted in the US East region:

   SPEECH_TO_TEXT_IAM_APIKEY=X4rbi8vwZmKpXfowaS3GAsA7vdy17Qh7km5D6EzKLHL2
   SPEECH_TO_TEXT_URL=https://api.us-east.speech-to-text.watson.cloud.ibm.com
   

Running locally

1. Install the dependencies

   npm install
   

Get the data

2. Run the rcv.js to get RingCentral Video Meeting Data

   node rcv.js
   

3. Run the rc-call-recording.js to get RingCentral Call Data

   node rc-call-recordings.js
   

4. Run the AI Webapp

   npm start
   

5. View the application in a browser at localhost:3000

Credits

IBM Watson orignal Sppech to Text App and resources

• IBM Demo: https://speech-to-text-demo.ng.bluemix.net
• Speech to Text AI API Docs : https://cloud.ibm.com/apidocs/speech-to-text

Support/Help

• Feel free to open pull request with any questions/issues/errors
• Please ask question, share comments related to this demo on RingCentral Developer Forum