How do I make a Whisper ASMR Pitch?

Close-mic whispered pitch over macro b-roll. Weirdly high retention; owns late-night scroll.

model⚡ Seedance 2.0 + ElevenLabs
agency would charge$1,200
you'll spend~$4
runs onTikTok

THE REFERENCE — MICHELOB ULTRA — SUPER BOWL ASMR

THE PROMPT — COPY, SWAP THE BRACKETS, RUN

VIDEO PROMPT (Seedance 2.0, no dialogue): Vertical 9:16, 12 seconds, seamless loop. SUBJECT: extreme macro shots of [PRODUCT — e.g. "a matte-black insulated tumbler"]. LOCATION: dark tabletop studio — black velvet surface, faint haze in the air, a single water droplet beside the product. LIGHT: one hard spotlight from top-right carving a rim highlight; everything else falls to black. CAMERA: macro lens feel, razor-thin depth of field, all movement at half speed.

BEATS —
0.0–1.5s: A fingertip enters frame and taps the product twice — slow, deliberate (the tap in darkness is the hook).
1.5–5.0s: Creeping macro push across the surface texture; a specular highlight slides along one edge.
5.0–9.0s: A fingernail traces a seam; slow rack focus from the label area to the water droplet.
9.0–12.0s: Pull back, then push in, landing on the exact framing of frame one for a clean loop.

No captions baked in, no watermark, no brand marks beyond the product's own label, five fingers.

AUDIO (ElevenLabs, separate pass): whisper voice, young female, extremely close-mic, breaths and mouth clicks left in, 60 percent pacing: "hey... you're still scrolling... so real quick... [PRODUCT]... it [CORE BENEFIT — e.g. "keeps ice frozen for 36 hours"]... that's it... that's the whole thing... okay... back to your scroll..." Layer tap and nail-trace foley under the VO in CapCut, synced to the visuals.

HOW TO RUN IT

  1. Generate video and audio separately, marry in CapCut
  2. Keep total under 12s
  3. Caption in tiny lowercase text, centered
browse all 101 ad styles →