---
url: https://superduperai.co/library/storytime-with-a-twist-format/
title: Storytime with a twist: a 60-second told story with the turn at 40 seconds
description: The talking-to-camera story that holds because the opener promises a turn and the turn arrives late. The prompt plans the beats, the b-roll inserts and the one line of text per beat.
category: formats
platforms: [tiktok, reels, shorts, youtube]
works_with: [claude-code, claude, chatgpt, any-llm]
tags: [storytime, talking-head, narrative, tiktok, youtube, trust, script]
difficulty: intermediate
verified: false
created: 2026-11-01
---

# Storytime with a twist: a 60-second told story with the turn at 40 seconds

The talking-to-camera story that holds because the opener promises a turn and the turn arrives late. The prompt plans the beats, the b-roll inserts and the one line of text per beat.

## Prompt

```text
Write a storytime script about {the story} for {topic or channel}. The format works because the first line promises an ending the viewer cannot guess, and the story is told in the present tense so it feels like it is happening now. The twist is a fact that reframes everything before it, not a joke.

Structure (60 s, 9:16, face to camera with 3-4 b-roll inserts):
- 0-4 s OPEN: state the ending without explaining it ('so that is how I ended up {outcome}'). On-screen: the outcome in 4-6 words.
- 4-15 s SETUP: where, when, what you wanted. Present tense. One b-roll insert of the place or the thing.
- 15-40 s BUILD: 3 beats, each ending on a small complication. One b-roll insert per beat, 2-3 s, over the line that mentions the thing it shows.
- 40-48 s TWIST: the fact that reframes the setup. Said plainly, no music swell. On-screen: the twist in 3-5 words.
- 48-57 s LANDING: what it meant, one sentence; what you do differently now, one sentence.
- 57-60 s CLOSE: 'Has this happened to you?' Comment ask. No download ask; if a product is in the story it is a prop in the b-roll, named once.

Word budget: 150-165 words at 165 wpm. On-screen text: one line at the open, one at the twist, nothing in between so the face carries it.

Return: the beat sheet (open, setup, 3 beats, twist, landing) as one line each, then the script as spoken prose with [b-roll: ...] markers, then the two on-screen lines.

The story, roughly: {what happened}
The twist: {the fact that changes it}
```

## Example

Story: a sourdough channel. Twist: the loaf that finally worked was the one I forgot about.

Beat sheet: Open: 'this is the loaf that fixed six months of bricks, and I did not bake it on purpose.' Setup: January, fourth starter, every loaf flat. Beat 1: a new recipe, a scale, a schedule on the fridge; flat. Beat 2: a colder kitchen, a longer proof; flat and sour. Beat 3: Friday night, shaped a loaf, got a call, forgot it on the counter until Sunday. Twist: baked it to throw it away; it rose, it had ears, the crumb was open. Landing: my proof was always too short, not my starter; now I proof overnight and bake in the morning.

Script: So this is the loaf that fixed six months of bricks, and I did not bake it on purpose. It's January. [b-roll: flat loaf on a board] Fourth starter, every loaf comes out like a frisbee...

...Sunday morning I find it, it has doubled, and I only bake it to prove to myself it's ruined. [b-roll: oven door, the loaf rising] It is not ruined. It has an ear. The crumb is open. [b-roll: cross-section]

The starter was fine the whole time. I was proofing for two hours in a kitchen that needed twelve. Now it goes in the fridge overnight and I bake with my coffee. Has this happened to you?

On-screen: 'the loaf I forgot about' at 0-4 s; '12 hours, not 2' at 40-48 s.

## Tips

- Say the ending first, but not the reason. 'How I ended up X' is the promise; 'because Y' is the twist.
- Record the whole story in one take, then cut the pauses. A story told in pieces sounds read.

## Works with

- Claude Code: Runs as a skill or a prompt in the terminal; can read your files and write the result.
- Claude: Paste into claude.ai or the Claude app.
- ChatGPT: Paste into ChatGPT; works with any GPT model.
- Any LLM: Model-agnostic: Gemini, Mistral, a local model.