LoRA for that general Jay Naylor style, trained using ~700 pics from 2017 up.

Use Model / Clip strength of 0.8, and negative prompts: speech bubbles, text bubbles etc to avoid artefacts.