How Stable Diffusion reads prompts
Many Stable Diffusion workflows use tag-style prompts, where order matters and earlier tags carry more weight. PromptLens converts its analysis of your image into ranked tags: subject and style first, then setting, lighting, color and camera.
You also get a negative prompt with common defects and unwanted elements, formatted with the standard “Negative prompt:” line so it can be split easily.
Tips
- Paste the first part into the positive prompt box and the line after “Negative prompt:” into the negative box.
- Use emphasis like (tag:1.2) sparingly on the details that matter most.
- Match your output resolution to the reference's aspect ratio for a similar composition.
Example Stable Diffusion prompt
portrait of an elderly fisherman, weathered skin, knitted wool hat, overcast harbor, soft diffused light, muted blue-gray palette, 85mm, shallow depth of field Negative prompt: blurry, low quality, watermark, text, distorted face