AI Parallel Lives

One clear portrait in JPG or PNG. Look for sharp focus, even lighting, and a face that is unobstructed and reasonably large in the frame. Sunglasses, heavy shadow, motion blur, or a tiny distant face leave the model far less identity to carry into each cell.

Yes. Fourteen life tags are one click away — astronaut, rock star, knight, chef, surgeon, firefighter, marine biologist, rally driver, ballerina, samurai, noir detective, fighter pilot, concert pianist, cyberpunk hacker — and you can type any other life in free text. Whole-grid set themes fill every cell at once.

Because a 3x3 grid splits the image into nine cells. At 1K each cell would be roughly 340 pixels wide, too small for a face to read. So 2K is the ungated floor and 4K is available on Pro. Generations cost 8 credits at 2K and 16 at 4K, identical on both engine tiers.

The template feeds your reference photo into every generation and locks the face, lighting direction and eye-line with explicit phrasing, so only wardrobe and setting change. Expect a strong match, and expect a busy 3x3 grid to drift a little more than a roomier 2x2.

3x3 gives nine lives and the classic 9 grid look people share. 2x3 gives six lives with more room per cell. 2x2 gives four large, detailed portraits and the tightest likeness. Choose by how much detail you want in each face rather than by how many lives sound impressive.

One person per run. The template is built around a single reference face so it can lock identity across every cell, and mixing people would weaken that lock. For a team, run each person separately and place the finished grids side by side.