I'm a product manager-turned-solo dev, and I'm building Promptsref, a curated Midjourney SREF library with full prompts, images and examples.
Promptsref helps you:
Quickly discover good SREF styles without trawling X/Discord/paid PDFs
Reuse full prompts + SREF to reliably reproduce a style
Explore by “vibe” (comic, cyberpunk, film still, etc.) instead of guessing keywords
Why I built it：
I started Promptsref after getting frustrated with Midjourney. As a beginner I kept seeing amazing styles, but they were:
1. buried in long X threads or Discord chats
2. shared as just a code or a partial prompt
3. almost impossible to reliably reproduce later
I wanted a place where I could save only the styles that consistently look good, with the exact prompt + SREF + example images in one place. That slowly turned into this project.
How it works：
For each SREF on Promptsref I store: the SREF code, the full prompt I used, example images (and sometimes video styles)
You can browse by tags and “style vibe”, copy a code you like, paste the full prompt into Midjourney, and usually get a similar result in a few generations. I update the library based on the styles I use in my own work and what the small community around the site seems to like.
Tech details:
Frontend: Next.js, deployed on Vercel
Backend: Supabase/Postgres to store SREFs, prompts, metadata and favorites
Things I’ve learned / had to work around:
Designing search for non-technical creatives is harder than just “add a search box”
Keeping SREFs useful when Midjourney updates models requires pruning and re-curating
People care a lot about consistency (series, brands, comics), not just one-off pretty images
What I’d love feedback on:
If you use Midjourney, I’d really appreciate feedback on:
Is the search / navigation understandable?
What information about a style do you wish you saw first?
Is there a better way to show series / collections of styles?
Any obvious technical improvements you’d make to the stack or data model?
Link: https://promptsref.com
Thanks for reading, and I’m happy to answer any questions about the implementation, data model, or the SREF curation process.
