Paste your image idea or positive prompt, select SDXL or SD 1.5 and identify the issue to exclude. Get separate editable positive and negative prompts with conflict checks. Copy each into its matching field in your image workflow.
Describe what you need, or load an example. You can edit and export the result.
Up to 3 previous results stay in this page session. Refreshing clears them.
A ceramic cup product photo with the brand word MOSS visible. Remove stray lettering, but keep MOSS.
Positive: A ceramic cup on a clean cream surface with the exact word “MOSS” visible on its side, soft light and sharp glaze detail. Negative: stray lettering, duplicated letters, extra logos, distorted cup rim. Check: Do not exclude all text because the brand word is required.
A garden portrait with visible hands. Reduce finger artifacts but keep flowers in the background.
Positive: A fictional adult in a garden, hands resting naturally on a table, flowers visible behind them. Negative: extra fingers, fused fingers, duplicated hands, disconnected limbs. Check: Preserve the requested flowers; do not add a blanket background exclusion. Results still depend on the model.
Choose the actual problem rather than collecting every negative keyword. The positive prompt describes the intended image; the negative prompt describes unwanted features. Review both together, especially when text, people or background objects are required.
This tool targets SDXL and Stable Diffusion 1.5 workflows with separate positive and negative fields. It does not supply embeddings, LoRAs or sampler settings and is not a FLUX negative-prompt tool. Change a few exclusions at a time and compare generated images.
