Products // 1 of 4 · Read

Grammata

In development

Every page, read: printed or handwritten, new or centuries old.

Grammata takes documents in from wherever they arrive (scans, phone photos, PDFs, mailboxes, archive boxes) and turns every page into clean, structured text: printed or handwritten, modern or historical, tables included, with a confidence for every word.

What goes in and what comes out

You send

  • Scans, PDFs and phone photos
  • Mailboxes and archive batches
  • Printed, handwritten or historical pages

You get

  • Text, line by line, with confidence
  • Tables rebuilt cell by cell
  • Page layout in reading order

Inside Grammata // 3 features

What it does.

Capture & quality checks

Every document, one front door.

Grammata takes documents in from wherever they arrive (uploads, scanners, phone photos, mailboxes and archive batches), checks that each capture is usable, splits multi-document files and removes duplicates before anything downstream spends compute on them.

OCR & handwriting

Print, handwriting, and the scripts others skip.

Grammata reads printed and handwritten text, line by line, with a confidence for every word. It is built for the hard cases: handwriting, degraded historical pages and low-resource scripts, where generic OCR quietly fails.

Layout & tables

Regions, reading order, every table cell.

Grammata turns a page into structure: titles, paragraphs, tables, figures, stamps and signatures, in reading order, with tables rebuilt row by row and form fields paired with their labels. Downstream products read structure, not a flat string.

See it work // illustrative

Grammata, on fictional documents.

The demo runs in your browser on fictional documents. It shows what Grammata returns; it is not production software and it measures nothing.

Pipeline · one document through the chain Illustrative
Supplier invoice Fictional through the chain
StageProductWhat it outputs
01 IngestGrammata · Capture & quality checks Scanned PDF, 1 page, capture quality 0.96: accepted
02 ReadGrammata · OCR & handwriting Printed French, Latin script; 3 words flagged below 0.90
03 LayoutGrammata · Layout & tables 13 regions in reading order; line-item table rebuilt as 5 rows × 4 columns
04 ClassifyKleis · Classification & routing Supplier invoice (0.97), routed to accounts payable
05 ExtractKleis · Field extraction 13 fields normalised; 1 below 0.90 held for review
06 EntitiesKleis · Names & entities 9 entities, 5 linked to known records
07 VerifyKleis · Consistency & ID checks 5 checks: 4 pass, 1 warning
08 AskSibyl · Questions & answers 3 answers with citations, 1 explicit abstention
09 SummariseSibyl · Summaries 3 sentences, each citing its regions
10 TranslateSibyl · Translation 2 regions, French → English, terminology kept
11 RememberMnemos · Document memory 4 claims written with provenance; matches a known supplier template
Illustrative: scripted outputs on fictional documents. Confidences and scores are examples, not measurements.

Try Grammata on your documents.

We start with a pilot on a sample of your own documents, measured before anything goes to production.