MidJourney finally released something people have been asking for forever…
a way to build your own visual style without prompts, guesswork or endless trial and error.
The new Style Creator (Beta) changes how you create visuals — and it might be their most important update yet.
🧩 What Is Style Creator?
It’s a new interactive tool that lets you create a personal or brand style by selecting images, not writing prompts.
No more typing:
“cinematic editorial soft lighting warm tones…”
Instead:
• You pick images you like
• MJ learns your taste
• You get a unique style code (like --sref 4669853)
• You reuse it for perfectly consistent visuals
It’s basically “training MJ” — without any training
How It Works 👇
- Start a style session
- MJ shows you a grid of styles
- Select the ones closest to what you want
- MJ refines and shows more variations
- After a few rounds, your aesthetic solidifies
- You get your personal style reference code
- Paste it into any prompt → consistent output every time
You can also share your styles or browse others in the community
Why This Matters 🔥
Brand consistency becomes effortless
You build a style once, then reuse it across campaigns, posts, ads or projects.
No prompt engineering needed
You can now express textures, moods, vibes and micro-details that words can’t describe.
Moodboards → obsolete
Style Creator is now your moodboard, style capture tool and aesthetic engine.
Perfect for teams and agencies
Share one style code and everyone produces the same look.
Your visual identity becomes reusable
Fans, clients or collaborators can generate in your style.
Pro Tips From Early Testing 🧠
• Early selections = broad exploration
• Later selections = precise refinement
• Your selection order heavily influences the final style
• Combine multiple style codes for interesting results
• Works even with short prompts
• It’s still beta — expect some quirks
Limitations To Keep In Mind ⚠️
• Strong styles can override literal prompt instructions
• Some old --sref codes behave differently now
• Grid sizes sometimes act weird (normal for beta)
• Results get better the more you iterate
Even so — this is a massive upgrade.
What This Means For AI Creators 🚀
We’re stepping into the era of personalized visual engines.
Not just generating an image… but generating your style on command.
Inside AI Visual Lab, I’ll be breaking down:
• How to build your own signature style
• How to blend styles for advanced looks
• How to use Style Creator for brand consistency
• Real examples + workflows
If you’ve already tried the new Style Creator, share your styles below — I want to see what you all are creating 👇