claude skills / Social & Content Skills / reel-scripts / en français

Reel scripts.

A vertical video is won or lost on the first frame. Claude writes the script the way it will be watched: with the sound off.

See what it produces
  • Instagram Reels
  • TikTok
  • YouTube Shorts
  • CapCut
  • Claude.ai
8formats that get reactions, from "guess first" to before and after
2.5spoken words per second at most
3hooks written: one chosen, two kept as variants to test

orders of magnitude, depending on your volume

the problem

What gets in the way today.

  • 01Your videos open with "Hi everyone" and a good share of viewers have already left.
  • 02You have a good idea, but no method to break it into 30 seconds, shot by shot.
  • 03The spoken text is too long for the runtime and the captions scroll faster than anyone can read.

input, output

You give this. You get this.

What you give Claude: the topic, the proof, what can be filmed and your voice

Business: Tidewell Plant Shop
Topic: check the soil before watering (the finger test from our care card)
Audience: indoor plant owners who overwater
On camera: hands only
Can film: two matching pots on the work table, terracotta wall, daylight
Platforms: Instagram Reels and TikTok, 15 seconds
Voice: friendly, a bit nerdy, "we" for the shop, two emojis max

$ claude "Write a 15-second Reel that makes people guess which plant needs water."

Reel script, script extractexample

Reel: "Which one needs water?" | Instagram Reels and TikTok | 15 s

Format: Pause and guess, because the audience thinks they can tell by looking, and the answer is a test they can copy.

HOOKS

A | Two identical pots labeled 1 and 2 | "Which one needs water?" | 4/4

B | Hand holding a watering can, frozen | "Stop. Check this first." | 3/4, less playable

C | Close-up of dry topsoil | "This soil is lying to you." | 3/4, vague

Chosen: A. Variants: B, C.

SCRIPT

0 to 2 s | two pots, same plant, same light | slow push-in | ON SCREEN "Which one needs water?"

2 to 5 s | same shot | a "3, 2, 1" counter appears | "Guess before the answer."

5 to 8 s | finger into pot 1, damp soil | "1: damp at 5 cm. Wait." | SPOKEN "Damp at five centimeters. Wait."

8 to 11 s | finger into pot 2, dry soil | "2: dry. Water it." | SPOKEN "Dry. That one's thirsty."

11 to 13 s | water poured into pot 2 | "Check, then water."

13 to 15 s | back to the two pots, as in frame 1 | "Did you get it right?" | loop

Spoken words: 10 | about 4 s

CAPTION

Most watering mistakes happen before the watering can comes out.

Comment 1 or 2 before you watch again.

#houseplants #planttips #plantcare #indoorplants

COVER

"1 or 2?" on the two pots, terracotta wall behind for contrast with the previous green cover.

SHOT LIST

1. Two matching pots, terracotta wall, window light

2. Finger test pot 1 (watered an hour before)

3. Finger test pot 2 (left dry for a few days)

the method

What Claude does, in order.

  1. step 1One sentence, one format

    "In 15 seconds, the viewer sees this proof, remembers this, then takes this action." Then the format, picked from 8 and justified.

  2. step 2Three hooks

    The first frame's image and text, plus a spoken line under 2 seconds. Each hook is scored out of 4 and the best one is kept.

  3. step 3The timed script

    Shot by shot: what we see, the single movement, the on-screen text, the spoken words, the sound, the cut. The spoken text is counted and timed.

  4. step 4Ending, caption, cover

    A last frame that loops back to the first, a single action, a short caption, and cover text that reads well in the profile grid.

demo in French / click, try it

Your turn.

An interactive demo of the skill. It is in French, the pack works in any language: ask Claude to answer in English.

Two identical pots labeled 1 and 2, slow push-in

Which one needs water?

0:00

the film / 44 seconds, with sound

The pack film.

the pack film

In 8 seconds.

the guardrails

What the skill never does.

  • The first frame is the hook: no intro, no logo, and the on-screen text already says what the video is about.
  • The proof is real and shown on screen. If it is missing, the script marks it [NEEDS PROOF].
  • Every spoken word is captioned, and the on-screen text tells the story on its own, with the sound off.
  • No music, brand or person without the rights: trending sounds come from the platform's licensed library.

Questions

Do I have to appear on camera?

No. The script adapts to what you want to show: you on camera, your hands only, a screen, the product alone, or text over footage.

Why a looping ending?

When the last frame matches the first, the video restarts with no visible cut. Viewers watch again, and the loop carries the call to action.

Does the skill edit the video?

No. It delivers the timed script, the shot list by location and the on-screen text. Filming and editing happen in your own tool.

What goes in the caption?

One strong line, one action line ("Link in bio." or the word to comment) and 3 to 5 relevant hashtags. Nothing more, unless you write long captions.

the pack

In the pack: Social & Content Skills.