The character image doesn't fit in consistently with the rest of the images. Also the character arc sometimes quotes the input verbatim. It would be better if the tool can express the idea better rather than just quoting the input literally