aicontentskills TikTok / Reel Hook Generator
Overview
The first 1.5 seconds of a TikTok, Reel, or Short decide whether your video goes anywhere. Every platform's algorithm weights early watch-through more than any other signal — if viewers don't stop scrolling in the first 1–2 seconds, the rest of your video never gets shown to anyone.
This skill generates visual hooks optimized for that first critical moment: 1–3 second AI-video openings designed to make thumbs stop moving. It produces multiple hook concepts for the same topic, each with a ready-to-copy AI video prompt, so you can A/B test which opening lands best.
When to Use
- User is making a short-form vertical video (TikTok, Reel, Short) and needs an attention-grabbing opening
- User has good content in the body of their video but the hook is weak and engagement is suffering
- User wants to A/B test multiple hook concepts for the same video topic
- User is batch-producing content and needs multiple hook variations per topic
Do NOT use this skill for:
- Long-form video (YouTube 10+ minute content) — hook rules are different
- Horizontal/landscape content — this skill is vertical-first (9:16)
- Full multi-shot videos — this skill generates the opening only; use ai-video-storyboard for full storyboards
- Generic prompt enhancement — use ai-video-prompt-enhancer for single-shot general-purpose prompts
Workflow
Step 1 — Understand the Video
Ask one combined question:
What's the video about, and what's the payoff? Tell me: (1) the video topic (what the viewer will see overall), (2) the target audience, (3) the CTA or takeaway at the end, and (4) any brand/product you want to feature visually.
Accept partial answers. If the user only gives the topic, that's fine.
Step 2 — Identify Which Hook Type Fits
There are six proven hook types for short-form vertical video. Pick the 2–3 that match the video's topic and vibe:
Hook Type 1: The Sensory Trigger
A close-up of something with strong texture, movement, or sensory appeal — food, liquids, fire, bubbles, fabric, ice, steam. Works because the brain is wired to pay attention to high-detail tactile content.
Best for: Food, cooking, lifestyle, ASMR-adjacent content, product unboxing
Example hook prompt:
Extreme close-up overhead shot of thick golden honey slowly dripping from a wooden dipper onto a stack of fluffy pancakes, steam rising, warm natural morning light, shallow depth of field with blurred background, cinematic 1080p, 1.5 seconds, 9:16
Hook Type 2: The Disruption
An unexpected visual that breaks from scroll-feed norms — an object in the wrong place, scale mismatch, inverted colors, something starting mid-motion. Works because the brain stops to process unexpected patterns.
Best for: Surreal/creative content, brand storytelling, edgy marketing
Example hook prompt:
A giant goldfish swimming through a busy city crosswalk at eye level as pedestrians walk around it unphased, overcast daylight, wide angle 24mm, slight motion blur, surreal cinematic look, cinematic 1080p, 2 seconds, 9:16
Hook Type 3: The Transformation Tease
Shows the "before" or "in-progress" state of something that will transform — a blank canvas, an empty room, a seed, a raw ingredient. Works because viewers stay to see the "after."
Best for: DIY, tutorials, before/after content, building/creating videos
Example hook prompt:
Close-up of weathered hands slowly unwrapping a single vinyl record from worn brown paper on a wooden table, warm late afternoon window light, shallow depth of field, muted sepia color grade, cinematic 1080p, 2 seconds, 9:16
Hook Type 4: The Human Moment
A face — especially eyes, a subtle expression, or mid-reaction. Works because human faces hijack attention more reliably than any other visual stimulus.
Best for: Storytelling, testimonials, personal narratives, reaction content
Example hook prompt:
Medium close-up of a woman in her late twenties slowly turning her head toward camera with a barely-there knowing smile, soft window light from the left creating gentle rim light on her hair, shallow depth of field with bokeh background, cinematic 1080p, 2 seconds, 9:16
Hook Type 5: The Motion Hook
Starts mid-motion — a runner already running, water already pouring, a car already in frame. Works because the brain can't finish processing incomplete motion, so it stays to see the resolution.
Best for: Sports, fitness, action content, product demos with movement
Example hook prompt:
Low-angle tracking shot of a runner's feet mid-stride on wet pavement in slow motion, water droplets spraying up with each impact, golden hour rim light from behind, shallow depth of field, cinematic 1080p, 2 seconds, 9:16
Hook Type 6: The Scale Reveal
Starts tight and implies something much larger — a close-up of a part that will reveal to a whole, or a single element in a much bigger scene. Works because viewers stay to see the reveal.
Best for: Landscapes, architecture, epic/cinematic content, brand films
Example hook prompt:
Extreme close-up of a single dewdrop on a blade of grass, perfectly still, reflecting a tiny inverted mountain in its surface, shallow depth of field, soft morning light, photorealistic detail, cinematic 1080p, 1.5 seconds, 9:16
Step 3 — Generate 3 Hook Variants
Produce exactly three hook concepts for the user's video, each using a different hook type. This gives them A/B testing options without overwhelming them with choices.
For each hook, output:
## Hook Variant N: [Hook Type Name]
**Concept:** [One-sentence description of what the viewer sees]
**Why it works for this video:** [Why this hook type matches the video topic]
**Prompt to copy:**
> [Complete cinematic prompt, 40-70 words, ending with duration, 9:16, cinematic 1080p]
**Duration:** 1.5-3 seconds (use whichever fits the rhythm of your full video)
Step 4 — Add Pacing Guidance
After the three hooks, add a short section:
## Pacing notes
- **Hook duration:** 1.5-2 seconds is the sweet spot. Shorter feels jumpy. Longer loses people.
- **First visible frame matters:** Your hook's opening frame is what people see in the scroll feed before any motion. Make sure frame 1 of the prompt makes sense as a still.
- **Sound hits at ~0.5s:** Whatever audio accompanies the hook (beat drop, snare, voice, whoosh) should hit 0.3-0.6 seconds after the visual starts. This is when the viewer's brain fully engages.
- **Cut to body content at the end of the hook:** The hook should feel like it's going somewhere, not like a complete thought. Leave a micro-cliffhanger.
Step 5 — Offer Caption Overlay Suggestions (Optional)
If the user wants it, suggest 2–3 text overlay lines for the hook:
- Question hooks: "Wait... what is that?" / "How is this possible?"
- Reveal hooks: "You won't believe this" / "Watch until the end"
- Direct hooks: "POV: you finally" / "This one trick changed everything"
Caption overlays are platform-dependent — TikTok captions should be <8 words, bold sans-serif, bottom-center.
Critical Rules
- Always produce exactly 3 hook variants. More is decision fatigue; fewer is no A/B option.
- Use different hook types for each variant. Don't generate three "sensory trigger" hooks — mix sensory, human, motion, disruption, etc.
- Match hook type to video topic. Don't use a "human moment" hook for a landscape video. Don't use "sensory trigger" for a tech review.
- Every prompt ends with duration + 9:16 + cinematic 1080p. Non-negotiable — this is the technical signal for vertical short-form.
- Hooks are 1.5–3 seconds max. Longer isn't better. Shorter is fine.
- Frame 1 must work as a still. Describe what's visible on the opening frame explicitly, so the model commits to a good thumbnail.
License
MIT — use freely, commercial or personal.