Hyper-Realistic Cinematic Action Scene from Reference Image

A detailed prompt for transforming a subject from an uploaded reference image into a hyper-realistic cinematic action scene. The subject is depicted leaping from a glowing, neon-swirling dimensional portal onto a wet city crosswalk, emphasizing dynamic motion, dramatic lighting, and preserving the subject's exact likeness and features.

Prompt by @புல்லட் வண்டி 😈🧛View the original post ↗Published 2025-12-18CC BY 4.0

The prompt

Fill in the blanks — 3 placeholders

Edit any of these and the prompt below updates. Copy gives you the filled-in version.

Using the uploaded image as the primary visual reference, transform the subject into a hyper-realistic cinematic action scene. The man is leaping forward onto a city crosswalk while emerging from a glowing circular dimensional portal behind him. The portal emits intense swirling neon energy in electric blue, violet, and magenta, with sparks, lightning arcs, particle debris, and visible energy-distortion effects in the air.

The subject is captured mid-motion in a powerful running leap, expression focused and intense. His original clothing from the reference image is naturally adapted for dynamic motion, with realistic fabric flow, folds, and physics, without changing his identity. Footwear and outfit subtly reflect the neon portal light. Preserve the subject’s original facial structure, skin texture, hairstyle, beard (if present), and body proportions exactly as in the reference image.

The environment is a modern urban city street with cinematic perspective. The background is slightly motion-blurred to emphasize speed and action. The road surface appears wet, showing subtle reflections of the glowing portal and neon light streaks. Dramatic cinematic lighting, volumetric light rays, shallow depth of field, ultra-sharp details, high contrast, realistic physics, sci-fi atmosphere, professional action-movie poster composition, cinematic color grading, ultra-high resolution, 8K quality.

Generate asks for your photo, then opens the editor with the prompt already loaded.

Model
GPT Image 1.5
Media
Image
Language
English
Words
209
Poster / FlyerCyberpunk / Sci-FiCinematic / Film StillFashion ItemCityscape / StreetAbstract / Background

How to use this prompt

  1. 1Copy the prompt — fill in any placeholders first and the copy takes your version.
  2. 2Paste it into your image model, or press Generate to open it here with your own photo.
  3. 3Run it twice. Generation is not deterministic and the second pass is often the keeper.
  4. 4Finish it with the free tools: upscale, crop to the platform you need, compress before sharing.

Go beyond generating the image

Everything after the generation, on the same site.

Upscale to 4KRemove the backgroundCrop for a platformCompress before sharingAll AI apps

More from GPT Image 1.5

Same model, same dataset — GPT Image 1.5 prompts people have shared.

View all GPT Image 1.5 prompts →

All image categories

Use cases
Styles
Subjects