Proof · NicNac Systems
Proof Before Pitch: What We Published Before We Asked Anyone to Buy
Anyone can ask you to buy. Fewer will show you the score first — including the parts where they lose. This is the founders offer retold the way we believe an offer should be told: proof before pitch.
The short answer: we measured our document parser, NicNacParse, against the paid service LlamaParse on the same seven documents, with one fixed scorer and the criteria written before either system ran. NicNacParse passed 21 of 21 mechanical checks. LlamaParse passed 20 of 21. The full study — including the three places the paid service is still better — was published whole on this blog before this post was written.
Why publish the score at all?
Because a score you can check is worth more than a claim you have to trust. The study used the same documents for both systems, the same scorer for both outputs, and criteria fixed in advance — so neither side could move the goalposts after seeing the results. That discipline is the product, as much as the parser is. An owner deciding whether a tool belongs in their business deserves that kind of evidence, in the open, before any conversation about buying.
Where the paid service is still better — in print, from us
The published study names three gaps plainly. LlamaParse describes images instead of only marking them. It reads heavy scans much faster — about 2.4 seconds a page against about 68 seconds for our three-pass reading. And it merges wrapped table cells more cleanly. Those gaps are printed with closing plans, not footnotes. If your work lives in heavy scans, that speed difference matters today, and you should weigh it honestly — we would rather lose a sale on a true fact than win one on a hidden one.
What this means for the founders offer
Our founders post on this blog sets out the founding position in the open, at rates that lock while a membership is continuous. This post does not change a word of it. It simply puts the order right: first the measured work, published whole; then the offer, for the people the evidence persuades. Proof before pitch is not a campaign line for us. It is how the parser study, the AI-visibility test we ran on ourselves — recommended in 0 of 28 scored answers, also published — and every other score on this blog got here.
Read the study. Check our work on your own document in the live demo. Then decide.
Read the proof
The full parser study is on this blog, with the score and the gaps in the same place.
Read the studyOne list, measured notes only
New articles when they ship - for owners who would rather read scores than slogans.
Double opt-in - a confirmation email completes your signup. Unsubscribe any time; our suppression list is global and permanent.