There are many AI Image generators that are offering one-bit masking for enhancement and editing. Ideogram, for example, and Deepdreamgenerator even handles masking outside the model selection.
Venice’s text based infill is very hit or miss in comparison. Implementing this feature is really essential for speedy image editing on the user end, and also lowers computing overhead on the server end by a significant margin, since only a small section of an image needs to be processed.