Template

Heavy Rain

Upload an image to start your template-based generation.

More

Upload Image

Heavy Rain

Credits Required:320

Easily Recreate the Same Look with Heavy Rain

Copy Heavy Rain prompts in one click and quickly recreate the same look in your images and videos, without starting from scratch.

AI Prompt

"Generate an 8-second hyperrealistic video, with 4K HD cinematic texture, {camera shooting pixel parameters 1.6}, natural image noise level 3.0, {realism parameter 1.6}, smooth light-and-shadow transitions, and perfectly accurate human body proportions; the protagonist’s facial features, gender, temperament, hair color, hairstyle, makeup, and clothing style are 100% based on the facial characteristics from @image. 【8-Second Video】 Overall style: A realistic local TV station breaking-news live broadcast feel, with subtle handheld-shake effects, lens smudges from rain, cold gray stormy lighting, wet road surfaces reflecting car headlights and police lights, dense visible raindrops, and no exaggerated transitions throughout—maintaining the raw, unpolished quality of the original footage. Shot-by-shot timeline: 0–3 seconds: Opening medium shot, fixed camera The scene is identical to the original video: a colorful row of houses along a street in heavy rain, with large puddles filled with black muddy water in the middle of the road, scattered gravel and orange traffic cones by the roadside, and a white SUV parked in the distance. The protagonist, acting as a news anchor, wears only an ill-fitting bright yellow hooded raincoat (with the hood loosely pulled over their head), clutching a news microphone emblazoned with the station logo, and solemnly reports to the camera: (stuttering pauses) “The rainin is very …… uh, heagty (a stuttering gasp followed by the mispronounced ‘heh’ for ‘heavy’) heh Nere York (mispronouncing ‘New York’), in vew York (again mispronouncing ‘New York’), in this is martialized ereda (intending to say ‘affected area’ but completely garbled into meaningless syllables), where — another one is gome (mispronouncing ‘gone’), with a casual posture that bears no resemblance to a professional journalist. In the distance, an ambulance flashes its red and blue emergency lights, driving toward the camera from the far end of the street. 3–5 seconds: Slight follow‑shot, sudden plot twist Hearing the sound of a vehicle behind them, the protagonist suddenly turns to look at the approaching ambulance, their expression shifting from reporting to alertness; the camera subtly shakes along with the turn, recreating the authentic immediacy of on‑scene photojournalism. The ambulance doesn’t slow down at all, directly plowing through a large puddle on the road, sending up huge splashes of muddy water that drench both the protagonist and the camera, leaving mud specks and water streaks across the frame. 5–8 seconds: Close-up, reaction, and conclusion The camera quickly zooms in to a close-up of the protagonist: they turn back toward the camera, their face covered in droplets of mud, hair plastered into damp strands sticking to their forehead, eyes wide open and mouth stretched into an O‑shaped expression—a stunned, bewildered look after being splashed. Bottom‑left corner caption: Retains the blue-and-white “4 NEWS” station logo, with a white banner bearing black text: “The raining is very heavily in New York...”; at the very bottom of the caption strip, a small blue band displays gibberish: “SPOMS TV SEAIN DONN R GOONE”; the sequence ends with a freeze‑frame on the protagonist’s shocked close‑up. Detail requirements: Dense falling rain threads, fragmented reflections of lights on the slick road surface, diffused glow of the ambulance’s red and blue emergency lights through the rain curtain, watermarks on the raincoat, blurred water stains on the lens, and slight exposure fluctuations when vehicles pass—no excessive beautification, preserving the rough, authentic feel of an outdoor storm‑live broadcast.

GumaGuma

Guma is an AI video and image generator that turns text and images into cinematic videos in seconds, powered by leading models including Kling, Nano Banana, Seedance and more.

Language

© 2026 Guma. All rights reserved.