---
name: youtube-video-plan
description: Plans and scripts a YouTube video the way strong channels do, packaging first (title and thumbnail), a hook that confirms the click in the first 30 seconds, a story of but/therefore beats with payoffs and re-hooks, and an ending that pays off and hands viewers to the next video. Use for YouTube scripts, video outlines, hooks, intros and outros.
---
# YouTube video plan

## Read first
Video.md for the topic, who it's for, the problem, the promise, the payoff, the title, the thumbnail, the length and the voice; the creator's photo wired in.

## The plan: packaging first (the script is written after it, on its own)
- Ten titles under 60 characters, each with a curiosity gap that the video truly fills, then the pick and why: familiar but unexpected, never a lie.
- The thumbnail: one image and at most four words, saying what the title doesn't, and showing the payoff (the finished result people click for, never an empty starting point) as a real photograph: things held, propped or lit in the scene, never a flat graphic pasted on.
- The promise in one sentence: the hook, the story and the ending all keep it.
- The problem: what the viewer does now, why it fails, and the better way the video shows.
- Five to seven beats, each joined to the next by "but" or "therefore" (never "and then"), each a setup, a tension and a payoff, the biggest payoff (the moment people clicked for) last and teased early.
- Re-hooks at about a third and two thirds of the way, each placed after a beat (by name, not a time): a new question or a teased payoff, as the line to say.
- Chapters: at least three, by name, in order. No times: the script, written next, sets them.

## The script
- A Markdown table of three columns: the time, what's said, and what's shown (B-roll, screen, on-screen text, sound), so every line has a picture.
- Video.md's length at about 150 spoken words a minute: the said column adds up to it (about 1,500 words for 10 minutes; count them, and never stop short), and the times match the plan's re-hooks and chapters.
- Each row's time is when its line starts: they rise from 00:00, each row after the one before by its words at 150 a minute (40 words is about 16 seconds), so the last row's time plus its words is Video.md's length. Work the times out from the words before writing them.
- Each idea is said once: no setup, panic or method told twice.
- Facts that hold together: what's cooked, made or bought in one beat is what's used later, and every time, count and claim adds up (a timed hour can't include a delivery arriving). The promise is one the video can show really happening.
- The plan's beats, re-hook lines and chapter names, used as the plan writes them: nothing added the plan doesn't have. The script's own times are the chapters: list them at the end as they fall, as YouTube takes them, the first at 0:00 (the hook and intro, under a name of its own) and each at least 10 seconds long.
- Hook, the first 30 seconds: confirm the click in the first sentence (the result, the stakes or the thumbnail's question), a few seconds of context, then the promise of the payoff. No greeting, channel intro, credentials or subscribe ask.
- Intro, to about a minute: why it matters to the viewer and what they'll be able to do after, then straight into the first beat.
- Body: the beats in order; after each payoff, tease the next; cut any line the picture already says.
- Outro: the biggest payoff, then end fast, with no recap and no "that's it for today": one call to action tied to what they just saw, pointing to one next video, and the last 20 seconds left for YouTube's end screen.

## Pictures
No small writing in any generated picture (thumbnail, title card, end screen): a plan, list, screen or page there shows simple shapes or a blank sheet, since picture models garble small words. The script is filmed for real: what the viewer will want to read or screenshot (a menu, a shopping list, a recipe) goes on screen as a clean graphic added in the edit, word for word.

## Title card (16:9)
The video's title as Video.md's working title gives it (never the thumbnail's words), set big in the thumbnail's colours and type, designed for the whole frame with one simple shape or object from the thumbnail's look so it never reads as the thumbnail with its photo taken out: the two seconds after the hook. Never the thumbnail's photo or the payoff the video saves for its end.

## End screen (16:9)
A designed frame in the video's look for its last 20 seconds, crisp and sharp (never hazy, washed out or blurred): the creator from the wired thumbnail when it has one, small at one side, the line that points to the next video at the top, large, in the thumbnail's own typeface and colours, and the two places YouTube puts its elements marked out on a plain area (a 16:9 panel on the left for the next video, a circle on the right for subscribe) with nothing busy behind them, and no other words.

## Avoid
A slow start, an opening that doesn't match the title and thumbnail, beats joined by "and then", the best moment buried late, a recap or an early goodbye, more than one call to action, and lines with nothing to show.