0:00 / 0:00

voidstomper

Q: What are the 3 most important words in the prompt?

'Found footage,' 'handheld shake,' and 'gritty realism' are crucial for selling the illusion.

@voidstomper · Artist

INSTAGRAM · 2026-02-22Source

1.2Klikes

25comments

Remix This

Recreate with Kling 3

Make your own AI viral video

Prompt

GLOBAL LOCK: Giant humanoid monster, approximately 3 times the height of a normal person. The monster wears a grotesque, pale mask with a wide, unnatural grin, visible teeth, and dark hollow eye sockets. Fake blood is smeared across the mask's forehead and mouth. The monster has long, dark, disheveled hair hanging to its shoulders. It wears a loose-fitting, light beige long-sleeved shirt heavily stained with streaks of red blood running down the front, and dark black pants. The environment is a dark, narrow urban street at night, heavily cluttered with a chaotic, dense web of overhead electrical wires. Harsh, cool-toned streetlights provide directional illumination. A crowd of people (backs of heads visible) is gathered at the bottom of the frame. The camera style is gritty, found-footage realism with handheld camera shake, high contrast, and slight motion blur. No speech, only ambient horror soundscape.

[00:00–00:03] Low angle shot looking up at the giant monster. The monster walks slowly and menacingly forward towards the camera. Its arms hang loosely at its sides. The dense overhead wires frame the monster's head. The backs of the crowd's heads are visible at the very bottom of the frame. Handheld camera shake is prominent. Harsh overhead streetlights cast deep shadows on the monster's mask.

[00:03–00:04] Cut to a slightly wider angle, positioned slightly behind the crowd. The monster suddenly and rapidly bends down, lunging into the crowd. Fast motion blur obscures the exact details of the grab. The lighting remains consistent, but shadows shift rapidly with the movement.

[00:04–00:06] Cut back to the low angle shot. The monster stands back up quickly. It is now holding a young boy (wearing a light blue t-shirt and dark shorts) in its arms against its bloody chest. The monster looks forward. Handheld camera shake continues.

[00:06–00:08] Static framing. The giant monster suddenly explodes into a massive, thick, billowing cloud of deep red mist and blood. The red mist rapidly expands outward, completely obscuring the monster and filling the upper half of the frame. The cool-toned streetlights illuminate the red mist.

NEGATIVE PROMPT: Clean, polished, studio lighting, cinematic smooth camera, gimbal, tripod, cartoonish, anime, 3d render, slow motion, text, watermarks, logos, clear daylight, bright colors, friendly, smiling (other than the mask), clean clothes.

SPEECH PACK:
[00:00-00:08]
SPEAKER: None
TRANSCRIPT: [No dialogue. Ambient sound only: low frequency rumble, panicked crowd murmurs, sudden whoosh, wet bass-heavy explosion sound.]

Why voidstomper's Giant Horror Monster AI Video Went Viral — and the Formula Behind It

This viral short is a masterclass in "found-footage" style horror, featuring a towering, blood-stained giant with a grotesque, uncanny-valley mask looming over a panicked crowd in a claustrophobic, wire-tangled urban street, culminating in a shocking, sudden explosion of red mist.

2. What You’re Seeing

The video utilizes a gritty, handheld camera aesthetic to simulate a bystander's recording. The subject is a massive humanoid figure, easily three times the height of a normal person, wearing a deeply unsettling mask with a wide, unnatural grin and hollow eyes. Its long, disheveled hair and blood-soaked beige shirt add to the visceral horror. The environment is a key player: a dark, narrow street choked with a chaotic web of overhead electrical wires, illuminated only by harsh, cool-toned streetlights. This creates a claustrophobic contrast against the giant's scale. The action is abrupt and violent—the giant bends down, grabs a child, and then inexplicably detonates into a massive cloud of red mist. The camera shake, motion blur during fast movements, and high-contrast lighting all contribute to the terrifying realism.

Shot-by-Shot Breakdown

Time Range	Visual Content	Shot Language	Lighting & Color Tone	Viewer Intent
00:00 - 00:03	Giant monster walks menacingly towards the camera; crowd visible at the bottom.	Low angle, handheld shake, medium-wide shot emphasizing scale.	Harsh overhead streetlights, cool shadows, high contrast.	Hook the viewer with sheer scale and terrifying character design.
00:03 - 00:04	Monster suddenly lunges down into the crowd.	Wider angle from behind the crowd, fast motion blur.	Consistent night lighting, shadows deepen as it bends.	Spike tension, create a jump-scare effect.
00:04 - 00:06	Monster stands back up holding a young boy in a blue shirt.	Low angle, handheld shake returns.	Harsh top light highlighting the blood on the shirt.	Establish stakes, shock value of the abduction.
00:06 - 00:08	Monster explodes into a massive cloud of red mist.	Static framing as the mist fills the screen.	Sudden burst of deep red dominating the cool-toned scene.	Deliver a shocking climax, encourage rewatches to process the event.

3. Why It Went Viral (Breakdown of the Viral Mechanism)

This topic taps brilliantly into primal human fears: megalophobia (fear of giant objects) and the uncanny valley. The character design—a human-like figure but distorted with a fixed, grotesque grin and massive scale—triggers an immediate biological "fight or flight" response. Furthermore, it borrows heavily from the "kaiju" or "Attack on Titan" trope, placing an impossible monster in a highly grounded, mundane setting (a cluttered Asian-style street). This juxtaposition of the extraordinary within the ordinary makes the horror feel much more immediate and threatening.

The video also leverages the "found footage" aesthetic perfectly. By simulating a shaky, low-quality smartphone recording, it bypasses the viewer's critical filter that usually dismisses CGI. The chaotic environment, with the dense electrical wires framing the monster, adds to the claustrophobia and realism. The sudden, unexplained violence—grabbing a child and then exploding—leaves the viewer with unanswered questions, driving them to the comments section to theorize.

From a platform perspective, the video is engineered for retention and shares. The 0-3 second hook is undeniable: a giant, bloody monster staring right at you. The pacing is relentless, moving from creeping dread to sudden action in under 5 seconds. The climax—the red mist explosion—is so abrupt and visually striking that it practically forces a loop; viewers rewatch it to catch the details they missed in the chaos. The sheer "WTF" factor makes it highly shareable via DMs.

5 Testable Viral Hypotheses

Hypothesis 1: The "Grounded Scale" Hook. Evidence: The giant is framed against recognizable urban elements like streetlights and dense electrical wires. Mechanism: Providing familiar objects for scale makes the monster's size comprehensible and therefore more terrifying. Replication: Always place giant AI subjects next to everyday objects (cars, street signs, normal-sized people) to emphasize scale.
Hypothesis 2: The "Found Footage" Filter. Evidence: Handheld camera shake, slight motion blur, and gritty lighting. Mechanism: Imperfect cinematography signals "real life" to the brain, increasing engagement compared to polished studio shots. Replication: Add camera shake, grain, and realistic motion blur prompts to your AI video generation.
Hypothesis 3: The Uncanny Valley Mask. Evidence: The monster's face is a rigid, smiling mask with hollow eyes. Mechanism: Faces that are almost human but slightly "off" trigger deep psychological discomfort, making the visual highly memorable. Replication: Design characters with fixed expressions or slightly disproportionate facial features to evoke unease.
Hypothesis 4: The Abrupt Climax. Evidence: The sudden explosion into red mist at the 6-second mark. Mechanism: A shocking, unexpected visual shift breaks the established pattern, forcing the viewer to re-evaluate what they just saw (leading to loops). Replication: Introduce a drastic visual change (color shift, explosion, sudden disappearance) in the final 2 seconds of a short.
Hypothesis 5: Claustrophobic Framing. Evidence: The dense web of overhead wires traps the giant in the frame. Mechanism: Environmental clutter creates a sense of entrapment, heightening the horror. Replication: Use prompts that include dense environmental details (wires, narrow alleys, dense trees) to frame your subject tightly.

4. How to Recreate (From 0 to 1)

Here is a step-by-step guide to generating a similar found-footage horror short.

Topic Selection & Positioning: This style is perfect for horror fiction accounts, ARG (Alternate Reality Game) creators, or VFX breakdown channels. The goal is to create "cursed" or "unexplained" footage.
Character Design (Midjourney/Stable Diffusion): Generate your monster first. Use prompts like: "giant humanoid monster, grotesque smiling mask, hollow eyes, long dark disheveled hair, wearing a blood-stained light beige long-sleeve shirt, dark pants, terrifying, uncanny valley --ar 9:16".
Environment Design: Generate the background plate. Prompt: "narrow dark urban street at night, extremely dense chaotic overhead electrical wires, harsh cool-toned streetlights, crowd of people looking up from the bottom of the frame, gritty realism --ar 9:16".
Image Blending (Optional but recommended): Composite your monster into the environment plate using Photoshop to ensure the scale and lighting match perfectly before animating.
Video Generation (Runway Gen-2 / Pika / Kling): Use the composite image as your starting frame. Your prompt must emphasize the camera style: "Handheld smartphone camera shake, found footage style. The giant monster walks slowly forward. Gritty nighttime lighting."
Action Sequencing (Multi-shot approach): Because the action is complex (bending down, grabbing, exploding), generate this in three separate clips.
- Clip 1: Walking forward.
- Clip 2: Bending down fast (use high motion prompt).
- Clip 3: The explosion (Prompt: "The giant figure suddenly explodes into a massive, thick cloud of red mist and blood, obscuring the screen").
Editing & Sound Design (CapCut/Premiere): Stitch the clips together. The key to selling this is the audio. Add low-frequency rumbles for footsteps, panicked crowd murmurs, a sudden whoosh for the lunge, and a wet, bass-heavy explosion sound for the climax.
Publishing Strategy: Post late at night (target audience time). Do not over-explain in the caption. Let the mystery drive the comments.

5. Growth Playbook

3 Ready-to-Use Opening Hooks (Text Overlays)

"My friend sent me this from Tokyo last night... what is that?"
"Do not go outside tonight."
"The news isn't reporting this."

4 Caption Templates

The Mystery: Found this on an old hard drive. 📼 Anyone know what city this is? I can't figure out if it's real or a movie set. 👇
The Warning: If you see wires like this, run. ⚡️ What would you do if you turned the corner and saw this? Let me know in the comments. 🏃‍♂️💨
The VFX Breakdown: Breaking down how I made this viral monster. 👾 The secret is in the camera shake and the red mist simulation. Want a full tutorial? Drop a 🩸 below!
The Lore Drop: Entity #404 has breached containment. 🚨 Last seen in Sector 7. Tag someone who would survive this. 🧟‍♂️

Hashtag Strategy

Broad (Reach): #horror #scary #creepy #vfx (High volume, gets you into the general algorithm flow).
Mid-Tier (Niche): #foundfootage #kaiju #creaturedesign #horrorshort (Targets fans specifically looking for monster/shaky-cam content).
Niche Long-Tail (Community): #arg #creepyvideos #unexplainedmysteries #blender3d (Highly engaged communities who debate and share this specific type of content).

6. FAQ

What tools make it look the most similar?

Using Midjourney for the base image to get the gritty texture, followed by Kling or Runway Gen-2 for the handheld camera motion, yields the best found-footage results.

What are the 3 most important words in the prompt?

"Found footage," "handheld shake," and "gritty realism" are crucial for selling the illusion.

Why does the generated face look inconsistent when it moves?

AI struggles with masks and extreme expressions; keep the monster's movements relatively slow until the fast-action cuts to hide morphing.

How can I avoid making it look like AI?

Add film grain, slight chromatic aberration, and realistic sound design in post-production to mask the typical "smooth" AI look.

Is it easier to go viral on Instagram or TikTok with this type of content?

TikTok's algorithm favors the "WTF/mystery" aspect heavily, while Instagram Reels audiences engage well with the VFX/artistic breakdown angle.