NicNac Systems
NicNac Systems
This is where NicNac systems stand next to the rented models they replace — same tests, same scorer, whole score published. Wins, ties, and the rows where we're still behind. If it hasn't been measured, it doesn't get a number here. It gets an empty slot, labelled.
Every study here follows the same discipline: study the best rented tool, build our own, measure both against criteria written first — then publish the whole result.
Seven document classes — book, flyer, multi-column, form, scan, sheet music, long report. Criteria written before either system ran. One fixed mechanical scorer for both.
Methodology: criteria written before any run · one fixed mechanical scorer for both systems · whole results published, including the rows marked BEHIND. LlamaParse is still ahead on image descriptions, heavy-scan speed, and wrapped-cell tables — each with a closing plan on the full scorecard.
Study scheduled — NicNacRouter vs rented API stacks. No numbers until it's measured.
Empty slot · No numbers yetStudy scheduled — NicNacBook vs rented schedulers. No numbers until it's measured.
Empty slot · No numbers yetNo benchmark theatre. A study appears here only when it was run this way — and the whole result comes with it.
Every acceptance check is written down before either system runs. Nobody moves a goalpost after seeing a result — including us.
A single fixed, mechanical scorer judges both sides against the same sources. If the scorer is wrong, that gets published too.
Wins, ties, and losses are published together, on the same page, in the same type size. A scorecard that only lists wins isn't proof — it's advertising.
Unmeasured comparisons sit in labelled scheduled slots. No projections, no “expected” numbers, no borrowed benchmarks.
NicNacParse is one of the systems these studies are about. Start with a $10 Taste — or see all eight $10 Tastes — or talk it through with a person first: (904) 689-3066. A person answers — that’s rather the point.