Transform a script into natural, human-like narration and generate speech using the TTS app. Use when the user wants to convert a script or text into expressive spoken audio with pauses, gasps, sighs, and other dramatic effects.
Transform a script into expressive, human-like narration and run it through the TTS app.
$ARGUMENTS is either:
If it looks like a file path, read the file first.
Rewrite the script to sound like a skilled human narrator by applying these techniques:
... (ellipsis) for short pauses, — (em dash) for dramatic pauses, or [pause] for deliberate beats*gasp* or the word "gasp" naturally where surprise occurs*sigh* or weave "sigh" into the text naturally*chuckle*, , or where appropriatehahehum, uh, well... for natural hesitationAfter transforming the script:
--instructions flag tailored to the script's tone and content. Examples:
"Narrate warmly and conversationally, with natural pauses and emotional inflection. Vary your pacing — slow down for dramatic moments, speed up for excitement.""Read like an audiobook narrator. Use a calm, measured pace with slight dramatic flair. Let emotions come through naturally.""Deliver with energy and enthusiasm, like a storyteller captivating an audience. Use vocal variety — whisper for suspense, project for emphasis."python3 tts.py "<transformed text>" --model gpt-4o-mini-tts --voice <voice> --instructions "<instructions>"
coral (or use the user's preference if stated)--file with a temp file instead of inline textgpt-4o-mini-tts since it's the only model that supports instructions