hi Rahma i’m sorry for the delay of the response, i recommend this tutorial https://www.youtube.com/watch?v=esWPb4i4vyY&ab_channel=InnovateYourselfInnovateYourself , i linked my rasa bot to Angular interface , the logic was to record the voice on the interfaceusing RecordRTC and then forward it to the NLU pipeline in the backend , first the voice will be coverted to text using google api speech to text , then we’ll get a text reponse from google api , after that this response will be treated by rasa pipeline and generates a response , after that the response will be converted back to voice using TTS (text to speech) feature using a python package pyttsx3 or you can simply use gtts from google api … i hope these information helps and i hope you fix it asap Best of luck