AI Demos Research — the structured-intelligence platform. Every verdict on these pages opens to the execution behind it.
Structured Document Extraction v1LlamaParseaggregate-across-documentsanswer
LlamaParse

Aggregate across documents

untested

"What's the total" — right only if every document contributed. Document grain, cross-document. Recorded as the first test to cut if the set needs trimming.

reading this cell
Every assessment ever made, whatever standard it was reached under.
reading policy v1 · fingerprint 4f53cda18c2baa0c…

What we measured

Nothing was measured numerically for this assessment.

Every assessment

0 counting · 0 from captured bytes · 0 transcribed

RefVerdictRubricProvenanceGraded byCounts here 

Nobody has assessed this tool on this question. `untested` is what the page says, and it is a value rather than a blank.

Every verdict on these pages is openable to the execution behind it. Where an execution was reconstructed rather than captured, the page says so — a finding and a measurement are not the same claim.