10% off monthly plans with code NODE10, until August 31
One photo of an object is enough: the sneaker squishes like memory foam, in extreme close-up, with the compression sound generated inside the video. The full workflow is visible here without an account.
Real workflow, real results. Pan and zoom freely.
Use this workflowThe workflow is right there: the object photo, the Camera node locking the extreme close-up, the squish instruction, and the Kling V3 video with its sound. Pan, zoom, open the nodes.
At a glance
The squish effect fits in one image: a rigid object, a sneaker, a mug, a car, compresses like jelly under an invisible press, then snaps back into shape. It is one of the most reused effects on TikTok and Instagram right now, and it needs neither a shoot nor a real object: a photo, a motion instruction, and the material does the show.
What separates a hypnotic squish from a failed one is not the prompt, it is the framing. The effect lives on the contrast between a perfectly still camera and a material that deforms: as soon as the camera moves, the eye follows the movement and the transformation goes unnoticed. That is why the template wires a Camera node into the circuit, extreme close-up, 135mm, locked camera, cabled straight into the video. The guide directing the camera explains this mechanic in detail.
On the model side, Kling V3 renders the five seconds in 9:16 with the audio generated inside the video: the soft, wet compression sound is half the effect, a silent loop keeps nobody watching. To start from a generated object rather than a photo, Seedream 5 Pro produces the opening packshot, and the public calculator gives the exact cost before any signup.
Import the object photo
A RIGID object, tightly framed, on a plain background: sneaker, mug, bottle, figurine. Perceived rigidity is what makes the compression spectacular, a squashed cushion surprises no one.
Let the Camera node lock the shot
Extreme close-up, 135mm, f/2.8, static camera: the node is already set. Do not put camera movement back into the prompt, the still/deformed contrast is the effect.
Describe the compression
The instruction tells the story of the material: the sole compresses, the upper folds into bulges, then everything bounces back. Say « like memory foam » or « like jelly »: that material reference is what the model follows.
Generate with the sound
Audio stays on: Kling V3 generates the compression sound together with the image. Five seconds is enough, the object returns to its shape at the end and the loop restarts cleanly.
The same canvas, tuned for different needs: each variation is two or three lines of prompt away.
The scroll-stopper of an ad: your product squashes then bounces back into place, and the eye stops. Start from your existing packshot or build one with the product photo workflow, keep your brand's background, and add the price or the hook in the edit, never in the generation.
The format that built the trend: extreme close-up, compression sound, five-second loop. The object must return to its exact shape at the end of the clip, that is what makes the loop invisible. Publish in native 9:16: cropping a 16:9 eats the material, and the material is what people came to watch.
A logo rendered as a rigid 3D object, glass, chrome or thick plastic, squashing like a gel cushion: the most shareable rebrand there is. First generate the logo in volume with Seedream 5 Pro, then run it through the template's Video node.
A croissant glazed like glass, a cake that looks like ceramic: the squish plays on material confusion. The harder and glossier the starting photo looks, the more the compression surprises. The food photo workflow prepares the plate, the squish deforms it.
The effect grows with scale: a car compressing like a sponge is more hypnotic than a mug. Frame wide enough to keep the whole object in the shot, but keep the camera locked: on a big object the tracking-shot temptation is strong, and it kills the effect.
A cushion, a plush or a piece of clothing squashing is the expected behavior: no surprise, no thumb-stop. The effect lives on the contradiction between perceived material and movement, it needs something rigid.
A zoom or a dolly during the compression and the eye follows the camera instead of the material. The template's Camera node locks the shot: do not unplug it, and do not write movement back into the video prompt.
The compression sound is half the effect, and it is generated with the video so it is perfectly in sync. A silent loop with music added in the edit loses the ASMR quality that makes people watch to the end.
A uniform compression looks like a rendering error. The template's instruction breaks it down: the sole first, then the upper in bulges, then the bounce. That material chronology is what makes the deformation believable.
Anything perceived as hard: sneakers, mugs, cars, consoles, figurines, rigid packaging. The more incompressible the object seems, the more the deformation surprises. Already-soft materials, cushions, plushies, fabric, fall flat.
Generated inside the video: Kling V3 produces the audio at the same time as the image, so the compression sound lands exactly on the deformation. Nothing to sync afterwards.
Yes, that is the main use: replace the template photo with yours in the Media node and keep the rest as is. For a consistent series, the product photo workflow builds the packshot first, the squish animates it next.
Kling V3 is the template's choice: five seconds in 9:16, audio generated inside the video, and good material handling. Seedance 2.5 is the alternative for longer variants or chained shots.
Around a dollar for the five-second clip with audio, opening packshot included if you generate it. The calculator gives the exact workflow price before any signup, and every node shows its cost before you launch.
Technically yes, Kling V3 accepts real faces, but on a human the effect disturbs more than it amuses. The trend was built on objects, and that is where the algorithm recognizes it. For people, the fashion transition is the effect that fits.
Explicitly ask for the object to return to its exact shape at the end of the clip, as the template's instruction does. If the last frame differs slightly from the first, trim the final frames in the edit: the end bounce gives you margin.
Yes: add an Image node upstream of the Video node, describe the packshot, and wire its output in place of the Media node. That is the right route for an object that does not exist yet, a prototype or a logo in volume.
Squish effect: the object that squashes like jelly
Use this workflowFree account, no card required. The workflow opens pre-filled in your canvas.
Kling V3
Seedream 5 Pro