How do you make AI ASMR videos?
Pick one calm, satisfying scene type (cutting, pouring, soft textures, rain, slow close-ups), write five to eight slow scene prompts with the same lighting and look, and render them vertically with a soft, sparse narrator, gentle music and subtle captions. Be clear about the limit: AI images and motion give you the visuals, so the 'ASMR' comes from pacing, mood and sound choices rather than recorded tapping or crunching. Label realistic-looking footage as AI-made when you upload.
- Veedio has an AI ASMR Video Generator for slow, calm, satisfying short videos.
- A five-scene video with still images costs 13 Veedio credits; with AI motion on every scene, 68.
- Music options include No music, Lo-fi and AI score, and caption styles include Subtle and None.
- Veedio produces narrated, captioned video with music; it does not generate recorded tapping, crunching or whispering sounds.
- YouTube asks creators to disclose realistic altered or synthetic content when uploading.
What 'AI ASMR' can and cannot be
Real ASMR depends on recorded sound: close-microphone tapping, brushing, crinkling, whispering. An AI video tool that builds narrated, captioned shorts does not record those sounds, and it is better to say so than to promise them. What it can do well is the visual half of the genre: slow, satisfying, oddly perfect scenes that people watch to relax, such as a blade slicing through a glass-like fruit, honey pouring in a long ribbon, rain on a window or a soft close-up of fabric.
That makes two honest ways to use it. The first is a calm visual video with gentle music and very little speech, which suits sleep and background-viewing audiences. The second is a quiet narrated piece, such as a slow description, a calm story or a breathing prompt, over those visuals. Both work as Shorts. If you want real tapping or crunching sound, record it yourself and add it in an editor after export.
Four AI ASMR formats and how to set them up
| Format | What happens on screen | Sound in Veedio | Best length |
|---|---|---|---|
| Satisfying cuts | A blade slowly slices glass-like fruit, soap, or layered cake, one cut per scene | Lo-fi music, minimal narration | 30 seconds |
| Pour and flow | Honey, paint, sand or molten sugar pouring in one slow ribbon | AI score or Lo-fi, minimal narration | 30 seconds |
| Rain and cozy rooms | Rain on a window, a lamp in a quiet room, steam from a cup | Lo-fi plus a soft narrator speaking one calm line per scene | 1 minute |
| Sleep story | Slow scenes of a quiet place, narrated gently to the end | Brian or Lily narration, Subtle or no captions | 1 to 3 minutes |
How to make an AI ASMR video
- 1
Choose one format and one object
Pick a format from the table, then a single object or setting that will repeat: a glass apple, a jar of honey, a rainy window. A channel that shows one kind of thing is easier to recognise and to binge.
- 2
Write one fixed look sentence
Describe the light, surface and camera once: 'soft diffused light, matte pale-wood table, extreme close-up, shallow depth of field, slow and steady'. Paste that exact sentence at the end of every scene prompt so scenes feel like one piece.
- 3
Write five to eight slow scenes
Each scene is one small, satisfying action: the first cut, the second cut, the slice falling away. ASMR-style videos escalate gently. Keep the hardest-to-look-away moment for the last third.
- 4
Pick the look
Cinematic gives realistic close-ups, Claymation and 3D render give a softer toy-like feel. Use one look for the whole video. If it looks real, plan to label it as AI in step seven.
- 5
Choose sound and captions
For a sparse, loop-style video pick Lo-fi or AI score music, keep narration to a few words, and set captions to None. For a quiet narrated piece, choose a calm voice such as Brian (deep, resonant, comforting) or Lily (velvety, cinematic), keep each line short, and use Subtle captions.
- 6
Render and check pacing
Render vertically at 9:16. Use stills with a slow pan for most scenes, and AI motion only for the two or three actions that need movement; motion costs 13 credits per scene against 2 for stills.
- 7
Upload with the right labels
Upload through YouTube Studio as a Short. If the footage looks like something real that never happened, turn on the altered or synthetic content disclosure and say in the description that it is AI-generated. Post at a steady rhythm in the same format.
Worked example: 'The Glass Apple' (30-second quiet narrated cuts)
5 scenes · about 30 seconds
Scene 1
Watch closely. Nothing here is in a hurry.
Visual: Extreme close-up of a translucent glass-like red apple on a pale wooden table, soft diffused light, shallow depth of field, slow and steady.
On screen: Slow
Scene 2
A single blade rests on the glass apple.
Visual: A thin steel blade rests against the top of the apple, a tiny reflection running along the edge, soft diffused light, slow and steady.
On screen: Watch closely
Scene 3
Slowly, the skin parts without a sound.
Visual: The blade slowly presses in, the glass-like skin parts cleanly, a faint inner glow, soft diffused light, slow and steady.
On screen: Calm
Scene 4
The first slice slides away, clear and smooth.
Visual: The first slice tilts and slides away, edges perfectly smooth and clear, pale wooden table, soft diffused light, slow and steady.
On screen: One cut
Scene 5
Even, glowing slices. Stay a little longer.
Visual: The apple cut into even, glowing slices fanned on the table, a single warm highlight, soft diffused light, slow and steady.
On screen: Stay a while
Scene prompts you can adapt
Each names the object, the single action, the light and the look. Add your fixed look sentence to the end.
Honey ribbon
Extreme close-up of golden honey pouring from a wooden dipper in one long unbroken ribbon into a glass jar, warm backlight, shallow depth of field, cinematic.
Use this promptRain on glass
A rain-streaked window at night, soft blurred street lights behind it, slow droplets sliding down, quiet and calm, cinematic.
Use this promptSoap slice
A pastel block of soap being cut into thin curled ribbons by a steel blade, matte white surface, soft diffused light, extreme close-up, claymation.
Use this promptCozy cup
A ceramic cup of tea on a wooden sill, steam rising slowly, a lamp glowing softly behind, snow outside the window, cinematic.
Use this promptSand pour
Fine pale sand pouring through a narrow gap into a smooth cone, side light making every grain visible, extreme close-up, documentary.
Use this promptSoft fabric
A hand slowly smoothing a cream knitted blanket, visible stitches, soft window light, shallow depth of field, cinematic.
Use this prompt
Recommended settings
| Setting | Choice | Why |
|---|---|---|
| Aspect ratio | 9:16 | Fills the Shorts feed |
| Length | 30 seconds for cuts and pours, 1 to 3 minutes for sleep stories | Short loops replay; longer pieces suit relaxation |
| Look | Cinematic for realism, Claymation or 3D render for a soft toy-like feel | One look per channel |
| Voice | Brian (deep, resonant, comforting) or Lily (velvety, cinematic) | Calm and steady; skip energetic voices |
| Music | Lo-fi or AI score | Gives the video a sound bed without speech |
| Captions | Subtle, or None | Loud captions break the calm |
| Credits | 13 for five scenes with stills, 68 with motion on every scene | Use stills for most scenes to keep cost low |
Mistakes that make AI ASMR channels fail
- Promising recorded ASMR sounds the tool does not produce. Say it is a calm visual video with music, and add real sound yourself if you want it.
- Changing the lighting and surface between scenes. The calm comes from everything matching.
- Cutting too fast. If a scene lasts less than a few seconds, it stops feeling relaxing.
- Using bold, bouncing captions or an energetic voice. Match the sound and text to the mood.
- Uploading realistic-looking footage without the AI disclosure label that YouTube asks for.
- Posting many near-identical clips. Repetitive mass-produced content is a monetization risk; vary the object, the action and the setting.
FAQ
Questions people ask next
Can AI make real ASMR sounds?
Not with Veedio. It makes narrated, captioned video with music and cannot generate recorded tapping, crunching or whispering. Use it for the visuals and the mood, and add recorded sound in an editor if you need it.
Are AI ASMR videos allowed on YouTube?
Yes, but YouTube asks you to disclose realistic altered or synthetic content, and its monetization policies reject repetitive or mass-produced content. Check the current policies before you plan a channel.
How long should an AI ASMR video be?
Thirty seconds suits satisfying cuts and pours that people replay. A sleep or relaxation piece can run one to three minutes. Veedio offers 30 seconds, 1 minute, 90 seconds and 3 minutes.
How many credits does one cost?
A five-scene video with still images costs 13 credits, and 68 credits with AI motion on every scene. Using motion only on two or three scenes sits in between.
Should the video have a voice?
It depends on the format. Satisfying cuts and pours work well with music and only a few spoken words. Rain, cozy rooms and sleep stories suit one calm voice such as Brian or Lily, reading short lines slowly.
Can I make AI ASMR videos for sleep?
You can make slow, calm visual videos with a gentle narrator and soft music. They are entertainment, not a medical or therapeutic product, so avoid claims about treating sleep problems.
Sources, checked 3 November 2026:
- YouTube Help: Disclosing use of altered or synthetic content
- YouTube Help: YouTube channel monetization policies
Platforms change their rules, so check them before you rely on them.
Keep going
Related
- How do you make AI cat videos for YouTube Shorts?
- How do you make motivational videos with AI?
- How do you make history Shorts with AI?
- How do you turn a long YouTube video into Shorts?
- How do you make 3D animation videos with AI?
- How do you make anime videos with AI?
- How do you make an AI cartoon video?
- How do people make AI story videos?
- What AI is used to make AI animal videos?
- What AI is used to make AI cartoon videos?