Adding support for voice recognition and also improving the system prompt

[pengu2004]

Hi, added a feature to accept voice input from the user by using the built in speech recognition api found in common browsers.

      recognitionRef.current.continuous = false;
      recognitionRef.current.interimResults = false;
      recognitionRef.current.lang = 'en-US';

1. recognitionRef.current.continuous has been set to false so the mic automatically stops listening when the user stops talking.
2. recognitionRef.current.interimResults has been set to false to only display results after the module has stopped listening. It can be set to true according to the preference of the maintainer
3. recognitionRef.current.lang = 'en-US'; is the language that is recognised by the browser.

I have also added a mic icon which calls the above function and starts listening.

Have modified the system prompt to avoid answering questions related to coffee making.

[vercel]

@pengu2004 is attempting to deploy a commit to the Sreenath S's projects Team on Vercel.

A member of the Team first needs to authorize it.