I just published the first version of node-red-contrib-vad.
This is an integration of the node-vad implementation of webrtc vad.
This node allows you to perform voice activity detection on pcm audio buffers.
Simply send a stream of buffers through this node and it will attach a message property to each message stating if the buffer in that message had voice in it or not.
This can for example be used in conjunction with a microphone stream and a speech to text node to only send audio containing speech to be transcribed.
It can also be used to build a effective voice command recorder used together with a microphone node to determine when the command was finished.
As always im happy about feedback, feature requests and bug reports on the nodes repo:
Best regards Johannes
