Curiosity hook examples that survive the three-second scroll

You scroll past a hundred videos before you stop on one. It happens in under three seconds, before the creator has said anything you'd call a pitch. That's the whole fight for a short-form video: not the editing, not the lighting, not even the topic. The first line, spoken or on screen, decides whether anyone sees the rest.

Most creators write that line last, after the footage is shot, as an afterthought bolted onto a finished edit. That's backward. The hook is the only part of the video every viewer sees, which makes it the only part worth obsessing over before you film anything else.

Why curiosity beats volume

The instinct is to open loud: a scream, a jump cut, a "WAIT FOR IT" in bold caps. Loud gets attention for half a second and then the viewer's thumb moves anyway, because loud doesn't promise anything. It's just noise with good lighting.

Curiosity works differently. A good curiosity hook names a gap the viewer didn't know they had and promises to close it. "The gym isn't the problem, your first week is" stops a scroll that "5 tips for staying motivated at the gym" never would, because the first line argues with something the viewer already believes about themselves. Argument creates a small itch. Viewers watch a few more seconds just to see if they agree.

That's the mechanical difference between the two most common hook types people actually use:

• Contrarian hooks take the thing everyone assumes is good advice and tell the viewer to stop doing it. They work because reversing a widely-repeated tip reads as information, not as a sales pitch.
• Question hooks ask something the viewer has personally wondered but never had answered. "Why does everyone quit by January 20th?" works on anyone who has quit a resolution, which is most people.
• Mistake hooks tell the viewer they're doing something wrong, specifically, right now. "You're not lazy, you're doing week one wrong" reframes a feeling of failure as a fixable error, which is more watchable than either blame or encouragement alone.

None of these require a loud opening. They require a specific, slightly uncomfortable claim delivered in the first sentence.

The on-screen text has a different job than the voice

Short-form platforms autoplay with sound off more often than creators expect, so the spoken hook and the on-screen text aren't the same sentence doing the same job twice. The text has to work as a hook on its own, for someone who can't hear anything yet.

That's why a strong hook script separates the two: a spoken line built for rhythm and delivery, and a shorter on-screen line built to be read in under a second, mid-scroll, with no audio. "Stop going to the gym every day" spoken alongside "Day 1 is where you quit" on screen gives a muted viewer a reason to tap sound on, instead of relying on the voice to do all the work.

Picking the hook, not just writing one

The harder problem isn't writing one good hook. It's that most creators write one hook, convince themselves it's clever, and never compare it against an alternative. A mistake hook and a contrarian hook aimed at the same video can land very differently with the same audience, and there's no way to know which without having both in front of you.

What the tool returns

A Capafy agent called Hook Lab exists for exactly this gap: a creator describes the video, and it returns a set of hook options built from the patterns above, instead of one guess. Here's a real, published example of what that input and output look like.

You send: "Hooks for a 30s TikTok about why most people quit the gym in January — energetic, for beginners."

You get back (shortened):

• Positioning: "The gym isn't the problem — your first week is." You own the beginner-who-keeps-quitting angle.
• Hooks (5+, each with spoken line + on-screen text + type):

1. Contrarian — 🗣 "Stop going to the gym every day." / 🖥 "Day 1 is where you quit" 2. Question — 🗣 "Why does everyone quit by January 20th?" / 🖥 "The January 20th wall" 3. Mistake — 🗣 "You're not lazy. You're doing week one wrong." / 🖥 "Week 1 mistake"

• Recommended pick: #3, with one line on why it fits a beginner audience.
• Script: 0–3s hook · 3–10s the mistake · 10–22s the fix in 3 steps · 22–30s CTA.
• Caption: short caption + CTA + hashtags, plus one filming tip.

The output isn't a single clever line, it's a comparison set across hook types plus a pick with a reason attached, which is the step most creators skip entirely.

Try it: https://capafy.ai/agent/8123079349?ct=hooklab_blog