332photographs from a real, already-edited wedding
57 slocal ingest · 6.1 MB working set vs 3.5 GB
46narrative scenes grouped
<$1full editorial read · 127 images
~$70printed album · 190 photos, 29 spreads
Measured
Finished gallery → local ingest
The application read the downloaded Pixieset folder locally in the browser, derived a thumbnail for each file, and discarded the heavy original.
57 seconds
6.1 MB working set
Measured
Scene grouping
The folder held two distinct sessions. The main event grouped into 46 narrative scenes.
46 narrative scenes
Measured
Editorial read
For each scene, representative photographs went to a vision model with craft instructions: narrative chapter, editorial rating, and a concrete twelve-word reason. The system proposes; per-chapter quotas set the final count.
127 of 332 images
9 batches · under $1
Validated
Spread composition
The album assembles from 24 designs in five families. The sequence follows the narrative order and never repeats a template twice in a row.
24 layouts
5 families
Validated
Correction
A rejected photograph promotes the next-best from the same chapter. A scene can move chapters; another instant from the same take can replace the pick.
Swap · move · mount
Authored proof
Physical print
A complete album went to print through Mixam and arrived in about two weeks. Quality is documented as below ultra-premium.
~$70 vs $100–300
190 photos · 29 spreads
Next proof
What still needs to be tested with Erika
The numbers above describe machine work. Erika’s review, correcting what doesn’t convince her and approving the final version, has never been timed start to finish. Next: join the four pieces into one application, put it in her hands on a real wedding, and measure gallery to approved album.
The number this
case will publish
Measured · ingest and editorial read · Validated · spreads and correction · Authored proof · Mixam print
