The structure scholars and ingest pipelines both rely on.
Multi-column layout is preserved, marginalia stays bound to its paragraph, and footnotes resolve to their reference.
We can clean and structure your back content of books, newspapers, and manuscripts so you can make that content discoverable for anyone relying on your services.








Read study →"I'm an expert at reading our 19th-century cursive, but I'm consistently surprised at the correct extraction of even badly handwritten letters. This is lifechanging."
Historical scripts, foxed pages, period typography
Try in playground →D · 02 NewspapersBroadsheet, multi-column, microfilm scans
Try in playground →D · 03 ManuscriptsHandwriting, marginalia, multi-script incl. CJK & Arabic
Try in playground →D · 04 Sheet musicScore notation, lyrics, structural metadata
Try in playground →D · 05 Special collectionsMaps, ephemera, photographic plates, broadsides
Try in playground →D · 06 Internal catalogsCard catalogs, finding aids, accession registers
Try in playground →Multi-column layout is preserved, marginalia stays bound to its paragraph, and footnotes resolve to their reference.
Latin, Cyrillic, Greek, Arabic, and CJK run today, with early-modern and pre-modern scripts expanding on the published multilingual benchmark.
Unique holdings and restricted collections run inside your own network on our same leading models.
Start with an API key — nothing to host or operate.
Sign up →Run in-region, with no egress to US infrastructure.
Sign up →Runs inside your own AWS, GCP, or Azure account.
Talk to sales →Fully offline, the same model weights, dedicated support.
Talk to sales →Free tier with up to $20 in credits per month — no card required.