THE EPIPHANY THING IS HERE! 🔥 Magnific Mystic SREF! 🔥 Introducing what is possibly the best generative AI STYLE TRANSFER that humans have seen so far for 1k, 2k & 4k resolutions ✨ This shouldn't be possible but now you have it! Details, tutorials & HUGE surprise 🧵👇

If you can't wait, here's the link to that tool that you already love (we ❤️ u back!). SURPRISE: I might be fired for saying this 😂 👉 if you are new, DO NOT SUBSCRIBE YET. Wait a week until BF because there will be a big dscnt! LINK 👇👇👇

So, this (+ what's coming) is the EPIPHANY thing! I was so obsessed for months about how to improve the richness of styles in Mystic and then I found a way to implement it... BUT 3m from idea to production. We are slow, shit. But better late than never!

I haven't created this sweet alone, I've been working closely with a can't-disclose-the-name-but-karanva team at Freepik who provided me with tech & infra! Let's see some examples! 👈 Left: ref. img. 👉 Right: generation 'an explosive woman in a spacecraft' KABOOM!

That image got 120k views the other day. You felt a disturbance in the force sensing that there was a new generative toy, didn't you?

👈 Left: reference image 👉 Right: generation with the prompt 'a beautiful woman' Go Akira, go!

👈 Left: reference image 👉 Right: generation with the prompt 'a woman with a hat smoking' For this one I used 2k resolution 🔥

Can you imagine the cinematic-style generations you'll be able to create with this piece of cake?

👈 Left: reference image 👉 Right: generation with the prompt 'a man surfing' Yep, it works perfectly both for realistic/film/photography and illustration/anime/artworks images. WHATEVER.

Full parameters:

And it's super obedient to the prompt, especially increasing 'Adherence'. 'A dog inside a crystal bottle', yeah I know, poor boy!

👈 Left: reference image 👉 Right: generation with the prompt 'a dog in a spacecraft' He is a gooooooood boy! He can smell aliens from so far away!

Another version with full parameters so you can play with it:

OK, one more. This one is so cool! I will let you guess which one is the reference and which one is the generation 😎

Did you guess right?

What I find amazing is that it works SO WELL for photorealistic styles. Sometimes it's very, very impressive. 👈 Left: reference image 👉 Right: generation with the prompt 'a handsome man medium shot'

Yes. Directly in 2k or 4k resolution with the style of your choice. I have some photographer friends who have been testing it last week and they are so hyped! Check out the small details 🤯 And now imagine this for advertising, marketing, fashion, etc... Wow!

Ah! It can write! Maybe you will have to increase 'Adherence' (under 'Advanced' settings) et voila! Prompt: A typical Japanese village during a war, an abandoned tank, illustration, huge letters with the word 'SREF' With 4 different reference images:

Another example: Prompt: In a mystic forest, huge letters with the word 'javilopen' - Adherence: 100% (better for texts) - HDR: 30%

So yes, I have to confess. I've been using the Mystic SREF secretly in some of my viral content (like in the aliens and the 'interstellar' trailer 😜). But now this power is on your hands too. But remember: with great power comes great responsibility!

OK, enough! TIME FOR YOU TO CREATE! GO! GO! GO!

Some users are mixing up our old but cool 'Style Transfer' (image to image) with the new 'Style Reference' (text to image). Here’s where you can find the new feature:

This made me very happy 💕 Thank you!

Show me your creations, please! 🤤🙏🏻

Had to share this one 😅 Nighthawks by Edward Hopper and: "two teddy bears gangsters in an American cafe at night, fierce look, both in a table drinking wine"