Stunning Light Studio · Prompt tools

Visual Coincidence

Three unrelated concepts. One impossible image. Pick the collision, choose the ingredients, and let the engine invent the reason they belong together.

Best experienced on desktop or tablet. Mobile works too — expect a little more scrolling.

← Studio Guide & FAQ Stunning Light BRENTCORP

1. Image & Output Settings

Universal settings for the Studio prompt builders.

Target image AI

Same creative choices; model changes tweak instructions, not ingredients. Midjourney remains experimental.

Reference images

Use reference imagesAdd instructions for images you'll attach later in your own AI. No files are uploaded here.
Strict Feature LockPreserve defining features when provided; request coherent anatomy for people and creatures alike.
Your photos stay with you. This site only writes prompts. Attach any references when you paste the prompt into your image AI.

Image format

Output character

Choose one primary render style. The middle panel adds the collision ingredients without overriding your selected look.

Output quality

ON by default; adds a quality instruction to the prompt.

Special instructions optional

Keep it short. Surprise Me won't erase what you write.

2. Build Your Collision

Pick a collision framework, then load the three concept slots — an object, a creature, a world. Each choice is one ingredient; Random tells the image AI to decide. Use Surprise Me for a fresh collision, or build your own one choice at a time.

The three concepts one pick each
Ingredients bend the collision

Leave blank if you don't want words on the image. Type a short phrase if you want words in the art.

New here? Hit ✦ Surprise Me first — see what it can do, then tweak what you get.

Fresh collisions, never premade prompts — Surprise Me rolls a new one every click. Custom text and shared image settings stay put.

3. Your Model-Ready Prompt

Built one choice at a time with full creative guidance. Copy into your preferred image AI.

Creative title not printed on the image

Prompt

STALE — the engine failed to refresh. Do not copy this text. Hit Surprise Me or change a setting to try again.

Ready. Build your collision one choice at a time.

We create instructions, not images. Image models interpret prompts differently; reference fidelity, poster lettering and exact aspect ratios depend on the AI you use.