Image sources and usage

Image sources

The catalogue preserves paired image/code evidence and original links from sref.site, SrefHunt, srefdb / jtatum and Pixmind-io. Some previews load from their original sites. Missing prompts, models and authors remain unknown. Midjourney results have not been independently reproduced; image reuse rights require separate review. Legacy concept images remain available through existing links. Categories are mapped from source tags.

How to document style results

1. Manually find candidate codes and record full prompts, model versions and supported parameters.
2. Test four fixed subjects: portraits, products, spaces and nature.
3. Retest the primary use with at least two different seeds, keeping failed outputs too.
4. Keep original images, job IDs, dates and rights records before editorial review.
5. Retest after model updates and keep version history.

The current styles do not include these verification records and should not be treated as reproducible SREF recipes.