Real-time sentiment analysis of audio conversations is now possible! Using Nexmo's websocket APIs and IBM Watson's speech-to-text and sentiment analysis products, developers can get raw audio data from live phone calls and analyze it in real time.

In this workshop, we'll create a serverless application that will accomplish these feats using the free (or 'lite') tier of IBM Watson services. While our initial implementation will be hosted on Glitch, this scalable solution could act as a prototype for a product that monitors enterprise-level amounts of data, for example performing a live analysis of an entire call center.

How to connect Nexmo and IBM Watson through RESTful APIs and WebSockets

Doing real-time analysis of audio streams

Performing speech-to-text decompositions on raw audio data

Performing sentiment analysis on the resulting text

Real-Time Voice Applications with Glitch, Nexmo and IBM Watson