Copy & Logic
- Factual accuracy verified against source claims.
- Voice coherence and strict tone compliance.
- Zero hallucinated external citations.
Our review standard connects the template, inputs, and displayed output. A test record describes a
particular tool, version, date, and set of runs; it is not a promise that every generation will match.
Each template has an identifiable goal, required inputs, and output format.
Generate a photorealistic industrial consumer device hero shot with clean studio rim lighting.
8K UHD, centered composition, soft shadows, 50mm f/1.8 optical bokeh, RAW photographic capture --ar 16:9 --v 6.1 --s 250
Raw AI OutputMidjourney v6.1 Run #04 • Seed: 83921049
Published PreviewTone curve applied in Lightroom.
Audit Pill: Disclosed Touch-Ups: Color grading applied in Lightroom, 0% structural changes
Published examples identify the filled prompt and relevant references. Any post-generation changes are disclosed.
We review the result against the task: factual accuracy for copy, subject consistency for images, and continuity or motion constraints for video.
“Tested” means a recorded review exists for the listed configuration. “Needs retest” means that evidence should not be treated as current. A retired prompt keeps an explanation or points to an appropriate replacement.
Audit performed within the last 30 days. High reproducibility with recorded test seeds.
Base model weights or system prompts altered upstream. Recalibration queue scheduled.
Technique obsolete or syntax deprecated. Card retains archived notes with link to current replacement.
Use “Report an issue” on a prompt page. Include the tool, prompt version, and what failed, without sharing private information.
No. It means we have a documented result for the conditions shown.
Yes, if the page clearly explains the editing. A heavily edited image must not be presented as an untouched prompt result.