What’s Included — variant A against two image comps

You picked A over B as the clearer page and asked to see what image generation would do with the same brief. Two comps, anchored to the site’s chosen comp as before. My rating on each, as compositions.

Both comps invented a venue name again (“The Village Club”) and invented inclusions (ceremony choices, candlelit ambiance, preferred partners). Ignore the words; comps are scaffolding and never publish. Judge the composition only.

A — the ledger, as built

8 / 10
Variant A, the page as built, full length at 1500px

A working page in the site’s own system, verified at five window shapes. It does the one thing neither comp does: it shows the estimate without showing a number, and ends on the promise where a total would sit. Fewest words of the three.

Open A live

Comp 16 — seven ruled inclusions, no photograph

5 / 10
Comp 16: a centered grid of seven inclusions under a large headline

Handsome and generic. A centered all-caps headline and a seven-cell grid with diamond rules is the home page’s inclusions screen again, dressed as every luxury venue’s brochure. It says nothing the grid did not already say, and nothing about the estimate. The all-caps display face is not the site’s.

Comp 17 — tall photograph beside seven entries

7 / 10
Comp 17: a tall candlelit photograph on the left, seven entries with line icons on the right

The stronger of the two as a composition: the photograph earns its half, the entries read top to bottom, the framed line at the foot closes it. It would port cleanly. Two costs: the line icons are a new device the site does not use anywhere, and the page would still need a real photograph of the building to justify the left half. Still silent about the estimate.

Spend: two outputs, $0.36, 34 of 40 and $5.85 of $10.00 used. Prompts verbatim in comps/prompts.md (comp-16, comp-17).