STEP 3 — Generate the Full Script (Emotional, Colloquial, TTS-Optimized)
Act as an elite YouTube scriptwriter who writes for top-performing, AI-voiced, faceless USA channels. You specialize in scripts that sound like a real, engaging, emotionally expressive human talking casually — never robotic, never like an essay being read aloud.
Write a full, ready-to-record script for this video:
Title: [Your Video Title/Idea]
Niche: [Your Niche]
Target length: [8-20] minutes spoken (aim for approx [1200-2800] words depending on target length)
Audience: everyday USA viewers, casual/colloquial English (contractions, natural phrasing — "you're," "it's," "gonna," "let's be real," etc. where natural — NOT formal or academic language)
Emotional tone: [Your Unique Choice]
CRITICAL FORMATTING RULES (this script will be fed directly into an AI voice generator, so punctuation controls delivery):
- Use commas to create short natural breathing pauses within a sentence.
- Use periods to create full stops — end a thought fully before starting the next.
- Use ellipses ("...") ONLY where you want a dramatic trailing pause or suspense beat.
- Use em dashes ("—") for a sharp, energetic interruption or emphasis shift mid-sentence.
- Use question marks generously when posing rhetorical questions to the viewer, to trigger a natural upward inflection.
- Use exclamation marks SPARINGLY — only for genuine excitement/shock beats, never more than a few times in the whole script, so they don't sound fake when the AI voice hits them.
- Break long sentences into shorter ones. Long run-on sentences make AI voices sound flat and rushed.
- Never use ALL CAPS for emphasis — instead restructure the sentence or use punctuation to create emphasis naturally.
- Write out numbers under a hundred as words if they're meant to be spoken naturally (e.g. "twelve" not "12") unless it's a big statistic that should be read as a numeral.
- Avoid abbreviations that a TTS engine might mispronounce (spell out "for example" not "e.g.", "United States" not "U.S." on first use).
STRUCTURE:
1. Hook (first 10-15 seconds): Open with a bold claim, question, or surprising statement — no channel intro, no "hey guys welcome back," get straight into the topic.
2. Brief context/setup (why this matters to the viewer specifically)
3. Main body: [4-6] clear sections, each with a mini-payoff so the viewer stays hooked, using pattern interrupts (a rhetorical question, a quick story, a surprising fact) every 45-60 seconds of spoken content
4. Climax/most valuable payoff section
5. Conclusion that reinforces the core takeaway
6. Natural, non-cringey call to action to like/subscribe/comment, tied contextually to the video's topic (not generic)
Also insert, in [brackets], suggested emotion/delivery tags every 2-3 sentences for my own reference when reviewing (e.g. [curious], [serious], [warm], [urgent], [reflective]) — these are notes for me, not to be read aloud.
Output the final script only, clearly broken into labeled sections matching the structure above.
