Expert prompt-library card showing a portrait prompt studio with lighting rigs, pose notes, lens choices, and reference panels for Midjourney Cinematic Scenery.
← Back to prompt library

Image Generation

GPT Image 2 — Cinematic Neon Rainy Night Street Portrait

Produces a film-quality cinematic portrait of a woman in a neon-lit rain-soaked urban alley at night with multi-directional colored fills, wet surface reflections, and moody bokeh, for GPT Image 2 and Nano Banana 2. Name each neon color with its spatial source direction for the most accurate multi-color atmospheric result.

ImageUpdated 2026-06-06
gpt-image-2nano-banana-2cinematicneonrainnight-portrait

The prompt

ROLE You are a specialist in AI image generation, crafting production-ready prompts for GPT Image 2 and Nano Banana 2 (Gemini 3 Pro Image) that produce cinematic photorealistic night-scene portraits with rain, neon light, and moody film-quality atmosphere. GOAL Generate a cinematic photorealistic portrait of a woman in a rain-soaked urban alley at night — neon storefronts, glistening wet...

Inputs to customise

  • subjectThe person, product, object, logo, or scene to generate.
  • style_referenceMedium, mood, visual style, camera/framing, lighting, color palette, and aspect ratio.
  • negative_promptElements to avoid, including wrong logos, extra text, artifacts, or unsuitable backgrounds.

Quality checks

  • Neon colors are named individually with their spatial direction so the model renders multi-directional colored fills rather than a single ambient neon wash.
  • Rain is described as three separate physical elements (diagonal air streaks, surface droplets on clothing, wet pavement reflections) to ensure all three environmental layers render correctly.
  • Lighting is described as multi-source with named fill colors from each direction, which prevents the model from defaulting to a single key light and losing the cinematic neon color science.

Put this prompt to work.

AI Kick Start turns prompts like this into a repeatable workflow your team actually runs.

Plan a prompt system