So this is a limitation of the TTS model itself by google, which reads markdown symbols (like asterisks) normally like a part of the actual text, for example, if it is given (Text-To-Speech) it would NOT read it as text to speech, it would actually read it "asterisk, asterisk Text-To-Speech Asterisk, Asterisk".
Even though this isn't a problem with the app itself, I'd suggest maybe implementing a system that strips away markdown symbols (like the Asterisk, hash, underscore, greater than etc.....) when actually sending the text to the model, the markdown is still there, but what the TTS reads is the raw text.
So this is a limitation of the TTS model itself by google, which reads markdown symbols (like asterisks) normally like a part of the actual text, for example, if it is given (Text-To-Speech) it would NOT read it as text to speech, it would actually read it "asterisk, asterisk Text-To-Speech Asterisk, Asterisk".
Even though this isn't a problem with the app itself, I'd suggest maybe implementing a system that strips away markdown symbols (like the Asterisk, hash, underscore, greater than etc.....) when actually sending the text to the model, the markdown is still there, but what the TTS reads is the raw text.