BC

BityClips

Prompt pack

All prompts

5 prompts · Video hooks

AI prompts for video hooks and cold opens

You want openings that survive the first three seconds on any platform.

The hook is the only part of a video that every viewer sees, which makes it the highest-leverage thing to generate variants of. These prompts produce hooks by mechanism rather than by vibe, so you can test one variable at a time instead of guessing.

Anyone with good videos and poor click-throughBatch-testing openings before productionRescuing an underperforming upload

The prompts

1. Twelve hooks by mechanism

Claude or ChatGPT

Generating genuinely different options

My video is about [TOPIC] for [AUDIENCE]. The payoff is [WHAT THEY GET].

Write 12 opening lines, one per mechanism: contradiction, specific number, warning, confession, unanswerable question, visual instruction, group callout, false assumption, countdown, mid-action open, cost framing, and time framing.

Each under 12 words. Then rank by stopping power for a viewer who has never heard of me, and explain the top three.

In practice: Asking for one hook per mechanism prevents twelve rephrasings of the same idea.

2. Hook plus first thirty seconds

Claude or ChatGPT

Making sure the hook is actually paid off

Take this hook: [PASTE HOOK]. Write the 30 seconds that follow it.

The opening promise must be partially delivered within those 30 seconds — not deferred to the end — while opening a second loop that keeps the viewer going. No recap of what the video will cover. No channel introduction.

In practice: A hook that promises and then stalls is worse than a boring one.

3. Diagnosing a weak open

Claude or ChatGPT

Fixing an existing video

Here is my current opening: [PASTE FIRST 30 SECONDS]. Retention drops [X]% in the first 15 seconds.

Diagnose it against: does the first line make a promise, is the promise specific, does anything delay it, is there throat-clearing, does it assume context a new viewer lacks. Then rewrite the opening three times with three different fixes, and say which failure each one targets.

In practice: Give the model the actual retention number — it changes how aggressive the rewrite is.

4. Title, thumbnail and hook alignment

Claude or ChatGPT

Removing the mismatch that kills retention

My title is [TITLE], my thumbnail shows [DESCRIBE], and my hook is [PASTE HOOK].

Check alignment: do all three make the same promise, does the hook repeat the title verbatim (it should build on it, not restate it), and would a viewer who clicked for the thumbnail feel the hook delivers?

Then give me one revised set where all three point at the same promise from different angles.

In practice: Mismatch between packaging and open is a common invisible retention leak.

5. Platform-specific hook rewrite

Claude or ChatGPT

One idea, four placements

Rewrite this hook for YouTube long-form, YouTube Shorts, TikTok, and Instagram Reels: [PASTE HOOK].

Account for the differences: long-form allows a slightly longer build, Shorts and Reels need the payoff signposted in under 2 seconds, TikTok rewards a conversational and slightly unfinished tone. Explain what you changed for each and why.

In practice: Reposting the same open across placements wastes three of the four.

Variables to fill in

  • [WHAT THEY GET] — the payoff, stated plainly
  • [X]% — real retention data makes the diagnosis concrete
  • [TOPIC] — narrow beats broad for hook writing

What actually improves output

  • Generate hooks separately from the script. Bundled requests produce generic openings.
  • Paste your five best-performing openings as a benchmark before asking for new ones.
  • Test one mechanism at a time so the result tells you something reusable.

Mistakes that ruin the result

  • Restating the title as the first line — the viewer already read it.
  • Deferring the payoff to the end of the video and calling it suspense.
  • Any greeting, channel name, or 'before we start' in the first ten seconds.

Tools that finish the job

Where these prompts go next

A prompt produces the script, the shot list or the markup. These are the tools that turn that output into a finished video.

Opus Clip

Auto-repurpose long videos into viral short clips.

Starts at $19/mo

View tool profile →

VidIQ

YouTube growth toolkit with AI-powered keyword and competitor insights.

Free, paid from $7.50/mo

View tool profile →

TubeBuddy

YouTube SEO browser extension for channel growth and optimization.

Free, paid from $4.99/mo

View tool profile →

Submagic

Auto-captions with stylized text and emojis.

Starts at $20/mo

View tool profile →

Apply these prompts to a niche

More prompt packs