How to Make an AI Squish Video (Squishy Face Trend, 2026)

What You're Trying to Make

One selfie in, and out comes the viral squish: a glossy, wet, slightly translucent silicone version of your head resting on a marble counter — and then two soft hands come in from above and squish, knead, pinch and pull it like a soft jelly toy, cheeks bulging, a quick eye-squish, everything slowly wobbling back into shape. Under it all, close-mic'd ASMR squelch synced to every press.

It's the oddly-satisfying, rewatchable side of the 2026 AI-ASMR wave. This guide covers the whole thing: the exact look that makes it work, the prompt, the tools people use, and the one-tap way to make yours.

Fastest way — Face Squish on Starrd bakes the whole glossy-head-plus-ASMR squish into one tap: upload a selfie, generate, and the wet silicone look, the varied hand squishes, and the synced squelch come for free — no prompt to write. Want the full method first? Read on. ↓

The Look That Makes It Work

Three things separate a real squish clip from a flat one:

• Glossy, wet silicone from frame one. The head reads as a shiny, slightly translucent jelly/silicone material before anything touches it — strong wet sheen, bright highlights. If it looks like normal matte skin, the squish reads as fake.
• A head on a surface. The signature composition is just the head — no body — resting face-up on a marble counter or wood table, shot straight-on or slightly overhead, filling the frame. It frames the face like a satisfying object to be handled.
• Varied, real hand moves. The hands should do different things — one-cheek grab, asymmetric knead, an eye-squish, a pinch-and-pull, a cup-and-pucker — not the same mirrored pinch on repeat. Slow, jelly-like bounce-back, never a snap.

The Two Ways People Make It

Search "squish" and you'll find two routes:

• CapCut / filter templates — CapCut's "Squishy" and "AI Squish Me" templates, Pika's original "Squish It" effect, and the various web squish generators. You pick a preset, drop in a photo, and layer an ASMR sound clip on top. Fast, but you're on rails: the same fixed motion every time, and the audio is added separately rather than matched to each squeeze.
• A one-tap generator — upload a selfie and the glossy-head look, the varied hand squishes, and the synced ASMR are generated together in a single pass. That's the lane Starrd's Face Squish sits in — no preset motion, no separate sound to sync.

The Squish Recipe

If you're prompting it yourself on a model with native audio (Seedance 2.0 or Veo 3), this is the shape that works. Feed it your selfie.

Main subject: you (from your uploaded photo), rendered as a glossy, wet, shiny silicone/jelly HEAD — just the head, no body — resting face-up on a smooth light marble countertop. Keep the same recognizable face, with a strong wet glossy sheen throughout.Camera: locked-off, straight-on and slightly overhead, the face centered and filling the frame. No cuts.Action: two hands enter from above and squish the glossy silicone head in a varied sequence of DIFFERENT moves — never the same mirrored motion. One hand presses and jiggles a single cheek while the other steadies the head; then both knead asymmetrically (one cheek up, one down) into a lopsided pout; then a hand presses over the eyes for a beat — eyes squish shut, jelly eyeballs bulge and wobble, then bounce back; then fingers pinch a cheek and pull it out, letting it wobble back; then both hands cup and squish the head up into a pucker. Deep, smooth, glossy deformation with a slow jelly-like bounce-back, never a snap. Recognizable and cute, never grotesque.Audio: close-mic'd, binaural ASMR — wet, glossy jelly/silicone squelch with a subtle pop and a sticky, tacky release, synced to each squeeze and slow release. No music, no speech. Generate audio.

Step 1 — Pick the Photo

One clear, well-lit, front-facing selfie, no heavy filters. It's the identity anchor — the model keeps your face recognizable while turning it into the glossy silicone head, and it adapts to whoever you upload. Dark, blurry, or group photos are the main cause of a face that drifts.

Step 2 — Generate (Prompt It, or Tap Once)

Path A — prompt a raw model. Paste the recipe into Seedance 2.0 or Veo 3 with your selfie. Both generate synced audio in the same pass, which is why they're the go-to for ASMR. Expect to iterate — the failure mode is a face that isn't glossy enough or hands that repeat the same pinch, so push the wet-silicone and "varied moves" language.

Path B — tap the template. Face Squish is Path A with the prompt already written and the glossy-head look, varied squishes, and ASMR pre-baked. Upload the selfie, generate, done. 100 credits (about $5.99), no subscription, credits never expire.

Common Mistakes

1. Matte skin. If the head isn't glossy wet silicone from the start, the squish looks like a bug, not an effect. Gloss first.
2. Both hands doing the same pinch. Mirrored, identical squeezes read as a cheap filter. Vary the moves — grab, knead, eye-squish, pull, pucker.
3. A fast snap-back. Real squishies ooze back slowly. A rubber-band snap kills the satisfying feel.
4. Music over the ASMR. The squelch is the content. Keep it ambient and close-mic'd; leave trend audio to whoever reposts it.
5. A bad reference selfie. Blurry or dark input = a face that drifts. Spend your effort here.

Related Reading

• How to Make an AI Pet Spa Video — the ASMR category's pet cut: the top-down facial with cucumber eyes and a cream massage.
• Viral AI Video Trends (2026): The Monthly Roundup — every trend worth making this month, each with a one-tap template.
• Seedance 2.0: The Complete Guide — the model most creators credit for realistic reference-photo video and in-pass audio.
• Seedance 2.0 Prompt Guide — the framework behind prompts like the one above.