Make one hero with both of them in it, then make it the reference for everything else. The hero prompt writes both identities once:
For shots 02 to 08 I attached that hero as the only image reference. Every prompt opens with one line that names both people again: same faces, his white hair and brows, her silver bun, same spots and wrinkles, same clothes.
With two people, the clothes matter as much as the faces. They are what tells the model whose hand or shoulder is whose in the close-ups.
Every shot names the same room and the same camera, word for word:
Only the Shot: line changes between frames. That fixed opening is what makes eight separate generations read as one real shoot.
For the hero and the last frame, ask for empty navy space at the bottom so there is room for text later.
AI loves young, smooth faces. Two people in their 80s only look real when you name every sign of age, then give each close-up its own job:
Each shot line tells the model exactly where the texture should be sharp.
Every prompt ends with the same two lines:
The first line says what real skin has. The second refuses the defaults that make AI faces look waxy.
No lipstick matters here: it hides exactly the lines you want to show.
Nothing is baked into the generated images, so the text always goes on last.
QC zoomed in. Regenerate just that one frame when:
Keep the rest of the set as it is. The eight ready prompts are in the members pack below.
This guide is the full process. The ready-made pack lives inside AI Creative OS on Skool, next to every other prompt pack I post.
