Four image prompts with caption space baked in. Free to copy.
What you get here
A single still sells the meme: a bridge ablaze at golden hour, a silhouette watching it go, embers where text can live. These four image prompts are written for ChatGPT, Gemini and Midjourney, with negative space baked in for captions. Copy, fill the [brackets], render.
Copy-paste prompts
Golden Hour Bridge Burn Landscape60 words
Best for: The hero image of the whole trend.
A cinematic landscape photograph of an old wooden suspension bridge on fire at golden hour. Flames glow amber against a soft peach sky, embers hang in the air like dust in sunlight, and the river below mirrors the burn in broken orange streaks. Shot on 85mm, f/2.8, rich shadow detail, subtle film grain, story of endings, [4:5 or 16:9] crop.
Silhouette Portrait With Drifting Embers56 words
Best for: For profile banners and posters.
Backlit portrait of a [person description] standing on a riverbank at dusk, body in full silhouette, facing a distant burning bridge. Golden embers float between camera and subject, catching rim light on their outline. Moody orange and deep brown palette, 135mm compression, gentle haze, quiet expression of moving on, shot for a [profile banner or poster].
Macro Ember Texture57 words
Best for: The background every text overlay wants.
Extreme macro photograph of glowing embers and cracked charcoal, orange light pulsing through black ash like a city seen from the air at night. Wisps of smoke catch the glow in the upper third, leaving clean dark space for overlay text. High contrast, sharp detail, shallow macro depth of field, textured background for [your quote or announcement].
Molten Letters Title Shot57 words
Best for: Announcements with actual heat.
A weathered wooden sign hanging over a river, its carved letters spelling '[your two or three words]' beginning to smolder and glow like cooling lava. Tiny embers peel off the letters and rise into blue-hour twilight, smoke curling around rusted chains. Photoreal, shallow focus on the sign, dark moody grade, space below for a [caption or handle].
How to use
Copy a prompt and fill the [brackets] with your scene.
Paste into an image tool — ChatGPT, Gemini or Midjourney.
Generate 2-3 variants; pick the one with the cleanest fire and darkest text zones.
Add captions in your editor and export at your platform's ratio.
Tips
Append “no text in image” — models love to add fake captions to fire scenes.
Golden hour and dusk skies read as “cinematic” faster than night shots do.
Ask for negative space where you plan captions; it beats fixing it in an editor.
Silhouettes solve faces: backlit figures stay on-model where portraits drift.
FAQ
Do these prompts work in ChatGPT and Gemini?
Yes — they're plain visual descriptions, so any image model can render them. Midjourney users should add aspect-ratio parameters and otherwise paste as-is.
Can I sell designs made with these prompts?
The prompts are yours to use freely; who owns the generated images depends on your tool's terms. Check your generator's commercial-use policy before selling anything.
How do I get space for text in the image?
Every prompt here bakes in negative space. If a render fills it up, append 'clean dark sky, no text, upper third empty' and generate again.
Why do hands and faces melt in fire scenes?
Put people in silhouette or behind smoke, as these prompts do. Fire-lit close-ups of faces are where image models still wobble most.