Audio input + speech-to-text recognition (either Tensorflow.js or premade package) - Githubissues

BenjaminBaffy / context-aware-personal-assistant

2 stars 0 forks source link

Audio input + speech-to-text recognition (either Tensorflow.js or premade package) #23

Closed MirrMurr closed 2 years ago

MirrMurr commented 2 years ago

To complete the User-Rasa pipeline we need voice input from user that can be converted to text, that can be sent to Rasa with the help of the BE.

What we need:

Microphone access
Speech-to-text conversion

Audio input solutions:

Built-in web api for Media Streams
p5.js sound library

Speech-to-text solutions:

Tensorflow.js + suitable algorithm (if we want to implement more deeply ourselves)
Existing package that does speech recognition