I kept running into the same problem while working with AI image and video tools:
Good prompts were everywhere, but they were hard to find again.
Some were saved in screenshots. Some were buried in browser tabs. Some came from image examples, prompt posts, social media threads, or generated results that looked useful but had no clean way to search them later.
So I started building soulg.xyz, a searchable visual gallery for AI prompt inspiration.
What it is
soulg.xyz is a gallery of 27,000+ AI visual prompt examples.
It includes references for:
- AI images
- AI videos
- posters
- character concepts
- product visuals
- PPT-style visuals
- design and layout inspiration
Each item is meant to be more than just a pretty image. You can open a detail page, look at the generated result, study the prompt structure, and reuse ideas for your own creative work.
Why visual search matters
A lot of prompt libraries are text-first.
That makes sense for code or writing prompts, but visual prompts are different. When someone searches for inspiration, they often do not know the exact words yet.
They might think:
- "cinematic portrait"
- "product poster"
- "cute dog"
- "luxury food poster"
- "anime character"
- "green visual style"
A normal keyword search can miss a lot of useful examples, especially when the original prompt is messy, multilingual, or written in a strange style.
So the goal is to make browsing and searching feel closer to how creators actually think: by visual direction, style, subject, and creative intent.
What you can do with it
You can use the gallery to:
- browse visual references
- search by simple words or ideas
- open prompt detail pages
- copy prompt structures
- compare different image styles
- find inspiration before writing your own prompt
The site is still evolving, but the basic idea is simple:
A visual reference library for people who create with AI image and video tools.
Public notes and selected examples
I also started a small public GitHub repository for selected examples and notes:
https://github.com/tyjean428/ai-visual-prompt-gallery
It is not the full product source code. It is more like a public explanation and example library that I can keep updating over time.
What I am improving next
The next things I want to keep improving are:
- better semantic search
- cleaner prompt detail pages
- better filtering by visual category
- more useful examples for designers and creators
- better organization for image and video inspiration
I would love feedback from people who use AI image tools, video tools, design tools, or prompt libraries.
What would make a visual prompt gallery more useful for your workflow?
Top comments (0)