How I use DiffusionBee for endless creative inspiration

screenshot of diffusionbee site. I also like the source was a codepen

tl;dr — since downloading DiffusionBee, I’ve been popping in prompts and coming back to images that make me feel inspired. It’s been a fun way to explore visual styles, and I’d highly recommend trying it if you’ve got an m1 mac.

pre-req — have a Mac (there’s a Windows 64 Bit waitlist as of 24 march 2023)

1. download DiffusionBee (free)
2. input prompt
3. view results, get inspired
4. make tweaks to prompt, repeat process

Note: If you are more l337 H@x0r and want to run stable diffusion from a CLI, these instructions were solid, but I err’d out somewhere and decided I should stop frantically pasting in sudo commands before something broke-broke.

Where to find prompts?

krea.ai, dalle2 prompt book, and arthub.ai. find an image you like, copy the prompt and then start tweaking it.

Here’s a few places for inspiration:

• krea.ai (cool company, fun site to explore)
• arthub[.]ai — I removed this link because there’s now (9 June 2023) just a bunch of nsfw flooding the homepage. You can still find images with the prompts used, but use this site at your own risk and if you haven’t visited it in a while…things have changed.
• https://dallery.gallery/the-dalle-2-prompt-book/ (official book that you can download the PDF, has a ton of information around styling options and is worth reading through)

I originally started with the Dall•E prompt book, but I wasn’t really happy with the results I was getting. Instead of trying to guess what results would happen with word-combos, I went with a more visual-first approach.

Krea.ai & arthub.ai both let you see images and the prompts used to create them (I’m sure there’s others out there, but I was able to find a couple starting prompts and haven’t been back).

a few examples from the Dall•E pdf. the information is fun to read through

The Dall•E prompt book is a great resource, listing prompt ideas and how words impact results. I’d recommend starting with a visual-first approach to quickly explore styles that resonate, and then reference the book to have a greater understanding of what’s going on behind the scenes.

A couple results

some results that I’ve been staring at

there’s something about this image styling that I really enjoy, and the ♾️ possibilities of results is really fun

Conclusion

If you’re looking for some inspiration, and want to input words into a computer and have it magically generate a bunch of bizarre images, download DiffusionBee and start messing around.

If you’re stuck on what to put into the prompt, just google ‘design prompts’ and you should see something that catches your eye, copy that prompt and begin to iterate on it.

I’d recommend starting with four images, and a smaller image size (~256x256) so you can generate enough images quickly to see what your changes are impacting (and if you like it).

Once you find a phrase that is giving you interesting results, increase the image size & # of images. (It takes a few minutes on my m1 mac mini, which is why I start with a lower res test).

I was liking the earlier results from this prompt, so I upped the image size & num of images

I’ll figure out something to do with all of these generated images, but the process is fun and I don’t know if I’ll get tired of staring at these things.

I hope you get as inspired as I have.

Best,

Kelsy

these images can be downloaded here

I liked this round of images (768x512), feel free to download.

• https://www.dropbox.com/sh/gn71cz8161j8wb5/AABXB3u4r69Y5UQ77_wMk_NTa?dl=0

screenshot of my history from 24 March 2023. general vibe has stayed pretty consistent.