Project Walkthrough

The
Visual Prompt Builder

I was taking an AI video bootcamp and the technical vocabulary was relentless. Shot types. Lighting setups. Realism triggers. Camera angles. Midjourney parameters. There was no way I was going to remember all of this unless I was a full-time cinematographer doing it every day.

So I built a tool that does two things at once: gives me a place to store everything I was learning, and turns all of it into a structured prompt generator so I never have to memorize the terminology again. Plain English in. Professional AI image prompt out.

Build Prompt tab showing all collapsed parameter sections: Shot Type, Lighting, Mood, Visual Style, Realism Triggers, Camera and Lens, and more

Every parameter category in one place. Expand what you need, leave the rest collapsed.

Tab 1

Build Prompt

1. Describe your scene in plain English at the top.
2. Work through the parameter dropdowns below.

Each one holds an entire category of professional vocabulary. Far more than anyone could hold in their head...

Golden Hour Dutch Angle Photorealistic Overcast / Diffused Cinematic Film Extreme Close-up Chiaroscuro Studio Softbox Volumetric / God Rays Low Angle Dark / Moody Handheld / Shaky Realism Triggers Moonlight Shallow Depth of Field Candlelight Nostalgic / Vintage Overhead / Bird's Eye Film Grain Aspect Ratio Surreal / Abstract Backlit / Silhouette Dolly Zoom Subtle Rim Light Natural Daylight Concept Art Neon / Cyberpunk Crane Up Gritty / Raw Wide Shot Soft Window Light Drone / Aerial Epic / Grand POV / First Person Color Grading Negative Prompts Blue Hour Orbit / Arc Fashion Editorial Medium Close-up Whimsical / Playful Realistic Shadows Texture & Detail Pull Back / Reveal Intimate / Personal High Angle Documentary Slow Push In

…

Click Generate and Gemini assembles everything into a structured prompt ready to paste into Midjourney, DALL-E, Kling, Veo 3, Higgsfield, or Gemini.

Build Prompt tab with Shot Type and Lighting sections expanded, showing clickable option tiles with labels and descriptions

Expand any section to read the plain-English description of what each option does. Click to enlarge.

The learning was built in

Every option explains itself.

This is the part that made it a learning tool, not just a utility. Every option has a plain-English description of what it actually does to an image. Click any tile to select it. The label tells you what it is. The description tells you what it does.

Shot Type - Extreme close-up to extreme wide, Dutch angle to bird's eye, and everything between
Lighting - Golden hour, chiaroscuro, volumetric rays, candlelight, and thirteen more
Mood and Atmosphere - Pick up to three to layer the feeling
Visual Style - Photorealistic, cinematic, editorial, and more
Realism Triggers - The specific phrases that push AI models toward photographic realism
Camera and Lens, Texture, Color and Tone, Film Grain, Negative Prompts, Aspect Ratio

I was learning what these terms meant by using the tool. The bootcamp was teaching the vocabulary. The tool was storing it and making it immediately usable at the same time.

Reverse Engineer tab with upload area for reference images, a field for what you like about them, and a field for what you want to create

Upload up to four reference images and describe what you like. Gemini builds a prompt to recreate that look.

Tab 2

Reverse Engineer

Find an image that has exactly the look you want but you have no idea how to describe it in a prompt. Upload it here and let Gemini figure out the vocabulary.

Upload up to four reference images. Describe what you like about them - "the mold texture on the wood, the moody moonlight, the way the shadows fall." Say what you want to create. Gemini analyzes the visuals and builds a prompt that recreates that look for your subject.

Grid Builder tab for building a labeled reference grid for Nano Banana, with title input and photo upload area for up to 9 images

Drop in object photos, add labels, download as a single reference grid image.

Tab 3

Grid Builder

Nano Banana is an AI video tool that needs a labeled reference grid of your objects before it can work with them. This tab is a dedicated builder for exactly that.

Drop in up to nine photos, add labels, and it assembles them into a clean 3x3 grid image you can download and drop straight into Nano Banana. A small tool solving a specific problem that kept coming up in the bootcamp workflow.

Reference tab expanded showing Course PDFs as chips, Boot Camp Classroom Links, and Camera Shot Types with icon grid

After Generation reference showing Creation Actions for Vary, Upscale, and Use, plus Animate Image options and Instructor Tips

Prompt Building reference (left) and After Generation cheat sheet (right). Click either to enlarge.

Tab 4

Reference Library

This is the storage half of the tool. Everything I was learning in the bootcamp went here. As I worked through each section of the course, I turned it into a PDF and saved it directly inside the app - so the reference material and the tool that uses it live in the same place.

Course PDFs - Cinematography reference, Nano Banana guide, Modern Cinematic Glossary, camera movements cheatsheet, and more - each one saved from the course as I completed it
Bootcamp Classroom Links - Direct links into the Skool course modules
Camera Shot Types - Visual icon grid with descriptions for each shot
Midjourney Parameters - Essentials, fine-tuning, reference, and version parameters
After Generation - Every post-generation action in Midjourney: Vary, Upscale, Rerun, Edit, Animate, and what each one actually does

The bootcamp content is inside the tool now. I didn't store it in a notebook I'd never open again. I don't have to log back into the course and hunt for where something was. Everything I learned is right here, inside the thing I use to apply it.

A year ago I would have built a custom GPT for this. I used to build GPTs for everything. Any course I took, any system I was learning, anything I wanted to be able to ask questions about later. Now I build apps.