On-device document OCR: pick an image of a document, KitDocReader (Unlimited-OCR,
~4.5 GB) turns it into structured markdown — tables survive as <table>/<tr>/<td> markup.
The whole ML surface is two calls:
let reader = try await KitDocReader(catalog: "unlimited-ocr")
let markdown = try await reader.read(imageAt: documentURL)xcodegen generate
open ReadDoc.xcodeprojRun on a Mac (the shipped bundle is macOS-only today). First launch downloads the model from the Hugging Face Hub; later launches load from cache.
swift run readdoc-cli --image sample.pngsample.png is a small rendered document (heading, paragraph, table) so the command works
out of the box.