AI ASMR Video Prompts: Glass Fruit, Soap and Crunch Sounds

AI ASMR became its own genre in mid-2025, when Google’s Veo 3 launched with native sound and clips of a knife slicing through fruit made of glass started spreading across TikTok and YouTube. The strawberry looked real, the cut sounded like glass, and the shards tinkled onto the board. Nobody could film that in real life, and that was exactly the point.

The format has grown since then: soap carving, crunchy honeycomb, crystal “ice” candy, tapping on impossible materials, peeling things that should not peel. What connects all of them is that the sound is the product. A beautiful clip with muddy audio is a failure in this niche, while an average-looking clip with a crisp, close, clean sound can still hold viewers.

These AI ASMR video prompts treat audio as the first line of the brief, not an afterthought. Each one names the trigger sound, describes the macro shot that produces it, and removes the things that ruin ASMR (music, talking, busy rooms).

Why ASMR Needs Its Own Prompt Approach

Most AI video prompts are visual first. ASMR flips the order, and that changes how you write:

• You must name the sound precisely. “Cutting sounds” is vague. “A sharp crystalline crack, then tiny shards tinkling onto a wooden board” is a sound the model can aim for. Onomatopoeia helps too: crunch, snap, ting, squish.
• You must control the mix. Models with native audio will happily add background music. Tell it what is loud and close, what is quiet underneath, and what should not exist at all.
• The action has to be slow. Fast cuts produce overlapping sounds that blur together. One deliberate cut, a pause, then the next cut gives each sound room.
• Macro framing sells the material. Extreme close-ups show the texture that the sound is supposed to belong to. When the visual and the sound match, the clip feels satisfying. When they do not, it feels off.

The AI ASMR Prompt Formula

[Extreme macro / close-up, static or very slow camera]
[Object + impossible or exaggerated material]
[Surface and background: plain, dark or soft, uncluttered]
[Slow action: one cut, pause, second cut]
[Trigger sound described precisely, loud and close]
[Secondary sound, quiet: room tone, shards settling]
[Lighting that shows texture: soft key light, rim light]
[no music, no voice, no narration]

Ready-to-Use AI ASMR Video Prompts

1. Glass Strawberry Slice (the Original Trend)

Extreme macro shot, static camera. A flawless strawberry made
entirely of translucent red glass rests on a dark walnut cutting
board. A polished chef's knife lowers slowly and slices it in
half. Each cut produces a sharp, bright crystalline "ting" and a
clean crack, followed by tiny glass shards tinkling onto the
wood. Soft warm key light, glossy reflections inside the glass.
Loud, close, crisp cutting sounds, faint quiet room tone. No
music, no voice.

Why it works: It names the exact sequence of sounds (ting, crack, tinkle) and keeps the camera still so nothing competes with them.

2. Glass Kiwi With Visible Seeds

Macro close-up, slow push-in. A kiwi made of clear green glass
with tiny black glass seeds visible inside sits on a white marble
board. A knife cuts three thin slices one at a time, pausing
between each. Every slice snaps with a crisp glassy crunch and
the slices tip over with a delicate clink. Cool soft daylight
from the side, sharp texture detail, shallow depth of field.
Close crisp foley, no music, no narration.

Why it works: Pauses between slices give each sound a clean start and end, which is what makes ASMR audio feel “clean.”

3. Soap Cube Carving

Top-down macro shot of hands in black nitrile gloves cutting a
pastel lavender soap block into small cubes with a thin steel
blade. The blade presses in slowly, then the cubes crumble and
fall onto a pile. Soft powdery crunch with each cut, a dry
scraping sound as the blade drags, gentle crumbling as cubes
tumble. Soft diffused overhead light, clean white surface. No
music, no voice, sounds close to the microphone.

Why it works: Soap carving is one of the oldest ASMR subgenres. “Powdery crunch” and “dry scraping” steer the model away from a wet or glassy sound.

4. Honeycomb Crunch

Extreme close-up of a slab of golden honeycomb toffee on a dark
slate board. A small wooden mallet taps it once and it cracks
into jagged pieces, then a hand breaks one piece in half toward
the camera. Loud airy crunch, sharp snaps, tiny crumbs pattering
onto the slate. Warm side light making the air bubbles inside
glow. Static camera, no music, no talking, crisp close foley.

Why it works: The airy bubbles inside honeycomb give the model a visual reason for a hollow, crispy crunch. For more food shots with appetite appeal, see our food and recipe prompts.

5. Frozen Crystal Grapes

Macro shot, static camera, dark background. A bunch of grapes
coated in a thin layer of frost and ice crystals hangs above a
glass bowl. A pair of scissors snips one grape free, it drops into
the bowl with a bright clink, then a fingertip crushes a second
grape and the ice shell crackles. Icy crackle, sharp clink, soft
fizz of melting frost. Cold blue rim light, sparkles on the ice.
No music, no voice.

Why it works: Three distinct sounds (snip, clink, crackle) happen at three distinct moments. Sequencing the sounds is more reliable than asking for “lots of icy sounds.”

6. Tapping on Gemstones

Close-up of a woman's hand with short polished nails tapping and
scratching slowly across a row of raw amethyst crystals on black
velvet. Each tap makes a bright hollow click, each scratch a soft
gritty rasp. Camera slowly slides left along the row. Moody single
spotlight from above, purple reflections, velvet absorbing the
light. Intimate close microphone sound, no music, no speech.

Why it works: Tapping and scratching are classic ASMR triggers. The slow camera slide adds motion without adding noise.

7. Peeling a Chocolate Shell Orange

Extreme macro shot of hands peeling an orange whose peel is made
of thin glossy dark chocolate. As the thumb presses in, the shell
cracks with a crisp snap and peels away in curved shards, showing
real juicy orange segments inside. Sharp snaps, faint juicy
squish, shards clattering lightly on a white plate. Soft warm
studio light, glossy chocolate reflections. No music, no voice.

Why it works: The contrast between a crisp outside and a wet inside gives two textures of sound in one short clip.

8. Sugar Glass Pane Fracture

Slow motion macro shot. A thin pane of clear sugar glass stands
upright on a wooden block. A fingertip presses its center until it
fractures into a spiderweb pattern, then collapses into sparkling
fragments. A tight crackling spread, then a soft cascading tinkle
as pieces fall. Backlit by soft white light so the cracks glow.
Close crisp foley, no music, no narration.

Why it works: Slow motion stretches the sound event, and backlighting makes the fracture pattern readable. Our slow motion prompt guide explains how to phrase it so the model does not just slow the camera.

Model-Specific Tips for AI ASMR

Veo 3.1: Where the trend started, and still strong on crisp sound effects. Describe sounds outright and use foreground/background language (“the crack cuts through a quiet room”). Veo 3.1 clips run up to 8 seconds, which suits one or two cuts. More on the Veo 3 prompt generator page.

Kling 3.0: Generates sound effects and ambience in the same pass as the video. Its longer clips (up to 15 seconds) let you do a slow sequence of four or five cuts in one take.

Seedance 2.0: Generates audio and video in the same pass and accepts an audio reference. If you already have a recorded cutting sound you like, you can feed it in as a reference to steer the texture. See our Seedance prompt generator page.

Hailuo: MiniMax’s H3 (Hailuo 3.0) adds native stereo audio. Older Hailuo versions are strong on material physics but output silent clips, so add foley in your editor.

Whatever the model, check the audio with headphones before posting. Most viewers of this niche watch with headphones, and small artifacts that are invisible on a phone speaker become obvious.

Common Mistakes With AI ASMR Prompts

• Leaving music allowed. If you do not say “no music,” expect some.
• Asking for too many actions. Five cuts in 8 seconds turns into a sound smear. Two or three cuts is plenty.
• Busy backgrounds. A kitchen full of objects invites kitchen noise. Use a plain board, velvet or a dark backdrop.
• Vague sound words. “Satisfying sound” means nothing to a model. Name the texture: glassy, powdery, airy, wet, hollow.
• Moving the camera fast. Whip pans and quick zooms break the calm. Keep it static or use a very slow push-in.

Good lighting prompts matter here too, because side light and rim light are what make materials look touchable.

How LzyPrompt Generates These

Type an idea like “cutting a glass watermelon ASMR” into LzyPrompt. Leave the dropdowns on Auto-detect, or set Shot Type to a close-up and choose a Lighting and Aesthetic that fit your channel. LzyPrompt returns 10 prompt variations, each with a scene title, a full prompt that includes the macro setup, the slow action beats and the sound cues, plus a ready-to-post caption.

The prompts are plain text, so they work in Veo, Kling, Seedance, Hailuo or any other model you use. For a series with a new material every day, export all 10 variations to CSV and generate them in one batch. Try LzyPrompt free for 7 days.