LongCat Image Edit performs precise text edits inside images. This post shows six real before and after examples that change only the target text while keeping the rest intact.
LongCat Image Edit is an instruction-driven image editor that handles both English and Chinese text. The model targets only the specified text while preserving layout, color, and nearby details. Model page: https://wiro.ai/models/meituan-longcat/LongCat-Image-Edit
Six base images were prepared and then edited with focused text-change prompts. Each edit changed a single line of text and kept other content unchanged. The tests used conservative guidance and fixed seeds for repeatability.
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
LongCat Image Edit handles targeted text changes with strong layout preservation. The model keeps fonts and nearby details stable in most cases. Small artifacts appear when the original text uses complex styling or tight spacing.