Markdown formatting disrupts text-to-speech output with unnatural formatting noise

An automated video generation system converted technical documentation into spoken audio using text-to-speech (TTS) engines. However, the inclusion of markdown formatting (e.g., bold, italics, code blocks) caused TTS systems to read formatting syntax as literal speech, disrupting the natural flow of the audio output.

Issue

TTS engines verbally read markdown formatting characters as literal speech, producing unnatural outputs like:

• ‘tool’ → “asterisk asterisk tool asterisk asterisk”
• “italic”