We are using RASA for a voicebot. We use speech to text which we submit to RASA. This gives us an intent. We were using non-streaming STT, meaning we process the text after we detect a silence in the speech, however we have now moved to real time STT. This gives us streaming text.
Question is whether we can process the streaming text in real time and generate an intent (or predicted intent) before the sentence finishes.