Script Writer

Written-for-reading and written-for-speaking are different crafts. Spoken content needs shorter sentences (the ear processes slower than the eye), more repetition (no rewinding to re-read), simpler vocabulary, and constant forward momentum. This skill produces scripts that sound natural when spoken — which means you can read them off a teleprompter without stumbling, and your listeners can follow on the first pass.

Every script is built around the same retention architecture. The cold open exists for one reason — to stop the 30-40% of viewers who leave in the first ten seconds — and it picks from four patterns (bold claim, result tease, story hook, contrarian question). Channel intros come after the cold open, never before. Each content section is structured Point → Proof → Punch → Bridge. Pattern interrupts get marked in the script every 60-90 seconds: tonal shifts, direct address, B-roll cues for video, vocal cues for podcast. CTAs are layered, not dumped at the end — a micro-CTA at 30%, a mid-roll at 50-60%, and the main ask in the last fifteen seconds when you have the most engaged remaining audience.

It writes the script exactly as it should be spoken. Contractions, phonetic spellings where natural, max 20 words per sentence (most should be 5-10), bold for vocal emphasis, [pause] and [beat] marked for delivery, stage directions in blockquotes. There's a short-form mode for TikTok/Reels/Shorts that compresses the whole architecture into 100-170 words — hook in the first 1-2 seconds, one idea only, no sections, and a deliberately abrupt ending to trigger algorithm-favoured rewatches.

Words-to-time conversion is built in: 120 wpm for podcast/educational pace, 150 for YouTube conversational, 170 for short-form. You always get an estimated runtime, three thumbnail/title concepts, suggested chapter markers, and notes on which clips would work as standalone shorts. It won't write "as I mentioned earlier" — that makes viewers feel they missed something. It re-states the point naturally instead.