🧰 Offline tools
Tools that run entirely on the cached Document Intelligence text in this machine's database — no Azure or GPT calls, nothing leaves the machine.
Library has 12 document(s); 12 have cached text and are searchable. Documents without cached text are silently skipped — run DI on them from the Library page if you need them included.
🔍 Tag lookup — which documents reference each tag
Tag list
Scope
Leave both filters empty to search every doc with cached text (12 docs). Use either or both filters to narrow — when both are set they intersect (a doc must match BOTH).
Preview
📥Download Excel| Document Number | Tag | Match Mode |
|---|
🔎 Attribute harvest — pull chosen attributes for a tag list, from cached text
Pick any attributes from your Attributes Library and paste a tag list; Lattice reads the value for each
straight from the cached document text. No GPT/Azure calls. Uses the Projects scope selected above.
Every result shows the source snippet so you can verify it.
Consolidate register is the deliverable: one row per register tag with its resolved Vendor and
PO, a decision queue for genuine conflicts, the evidence behind every value, and — in the same pass — whether each
tag was found exactly, only by ignoring separators, is OCR-corrupted, or isn't in the cache at all.
Harvest is a quick first-document pull; Audit sources drills into where
documents agree or contradict; Raw text dumps the cached text around barren tags.
(The Layout selector applies to Harvest only.)