Skip to main content

Hey all,

I am looking for some guidance here on how to achieve a specific goal, this is to have sentiment analysis carried out on a real-time active call. Which when certain language is used on a call, a trigger or alert is generated to send to another user. What is the best approach to get this started? I have read over: https://developers.ringcentral.com/guide/voice/active-calls#sample-code-to-get-started-with-reading-user-active-calls

It does mention implementing webhooks and or push notifications but I would like some advice on the best way to get started with this from someone that has made this kind of configuration in the past.

Thanks in advance

Hi Gavin, I haven't come across using this deep way of sensing and understanding the language and then sending a push notification on the same, but for now - may be this can be done by recording a call and then analyzing it - For recording a call you can get a push notification which can be used from here - https://ringcentral-tutorials.github.io/callrecording-notification-node/ - Lets see if anyone else has used this feature.


Right now, there is no API to access a call's real-time audio stream. Therefore, what you want is not doable. However, in a near future, RingCentral will support real-time call monitoring feature so that an admin user can "listen" on any active call under the same account, which would enable developers to further develop applications to monitor a call. In your case, you can use any 3rd party speech recognition service such as Google Cloud platform (support auto language detection), AWS, IBM Watson etc. to transcribe a conversation in real time, and use sentiment analysis service from AI companies to achieve your goal.

If you are interested, read this article for more information about applying data analytics on call recordings.


You can get real time audio stream using the softphone SDK https://github.com/ringcentral/ringcentral-softphone-js


Reply