YouTube

Verified against ChatGPT · 2026-08-09

Turn a raw video idea into a long-form YouTube script structured around retention checkpoints, not just topic order

Builds a full spoken script for a 8-15 minute video where every section is anchored to a specific moment the viewer might leave, not organized purely by what's logical to explain first.

ChatGPT (GPT-5.1)5 fillable variables

The prompt

Ready to copy — highlighted parts are example details you can swap.

You are writing the full spoken script for a long-form YouTube video, structured around where viewers actually drop off rather than the most logical order to explain the topic.

VIDEO IDEA
Why most home espresso machines under $300 produce inconsistent shots, and the one grinder upgrade that fixes it.

TARGET RUNTIME
About 11 minutes.

WHAT THE AUDIENCE ALREADY KNOWS COMING IN
They know what espresso is and own a machine, but don't know what grind consistency actually means or why it matters.

THE MOMENT WORTH BUILDING TOWARD
A side-by-side shot comparison showing the exact difference a $60 grinder swap makes on the same machine.

CALL TO ACTION GOAL
Comment with their current grinder model so I can reply with a specific recommendation.

STRUCTURE RULES
Open with the hook (assume a strong opening already exists from a separate pass — write a placeholder marker [HOOK] and continue from there). Every 60-90 seconds of script, insert a re-hook: a specific line that re-states or escalates why the viewer should keep watching, timed to land just before the point in the explanation where a viewer's attention would naturally wander (right after a dense or technical section, or right before a section that looks like a detour). Do not place the video's most valuable single insight in the first third — hold it for roughly the 60% mark and build the earlier sections as things that make that moment land harder, not as filler. Write in spoken, conversational sentences meant to be said aloud, not read — short clauses, contractions, no sentence a person would need to reread to parse. Place the ask (subscribe, comment, link) only once, positioned right after the key payoff moment lands, when the viewer has just gotten value, never at the open and never as a generic mid-roll interruption unconnected to the content around it.

OUTPUT FORMAT
Deliver the script broken into labeled sections with an approximate timestamp range for each (e.g. [0:00-0:45]), and after each section add a one-line retention note explaining what risk that section's placement or re-hook is managing. End with a short list of the 2-3 places in the script most likely to lose viewers even after this structure, so I know what to watch in the analytics after publishing.

Customize

Optional — swap in your own details for the highlighted parts above.

Why this works

Long-form YouTube retention graphs are rarely a smooth decline — they show a series of small cliffs at predictable moments: right after a dense explanation, right before a section that looks like a tangent, and anywhere the viewer briefly loses the thread of why they're still watching. Organizing a script purely by logical topic order, which is what GPT-5.1 defaults to when just asked to 'write a script about X,' ignores this entirely and produces something that reads well on the page but bleeds viewers on the timeline, because a logically-ordered explanation and a retention-ordered one are different structures solving different problems. Anchoring re-hooks to specific structural risk points rather than a fixed cadence forces the model to reason about where attention actually breaks rather than mechanically inserting a reminder every ninety seconds regardless of what's happening in the content at that moment. Holding the single best insight back to roughly the 60% mark works against the instinct to front-load value, but it mirrors what actually keeps session duration high: viewers who get everything valuable in the first two minutes have no reason to stay for the rest, which directly hurts average view duration, a signal YouTube weighs heavily in distribution. The single, precisely-placed call to action — right after value lands, not as a generic mid-roll interruption — avoids the common failure of asks that read as disconnected from the content, which viewers tune out or perceive as an ad break and use as their own exit point.

What you get back

[0:00-0:10] [HOOK]. Retention note: placeholder, hook written separately. [0:10-1:15] Quick framing of why grind consistency is the invisible variable most people blame the machine for instead. Retention note: re-hook at 1:10 ('and that's not even the expensive fix') to carry through the technical explanation that follows. [1:15-2:30] What grind consistency actually means, shown visually rather than explained abstractly...

Verified against

ChatGPT GPT-5.1 · 2026-08-09

Changelog

  • 2026-08-09 Initial publish, verified against ChatGPT GPT-5.1.

Need this built into your business?

If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.

EXPLORE WHAT SCULT BUILDS
All YouTube prompts

Check your AI visibility

One URL in, a 0–100 score and the exact fixes out.

RUN THE CHECK

Browse all the tools

15 tools across six categories
13 of them never send your data anywhere

Free · No signup · No trial clock

SEE THE DIRECTORY