At Zipphy, I worked on solving similar problems in on-prem environments — building an OCR + NLP + CV pipeline to generate spatial layouts and classify documents at scale.

One persistent challenge was generalizing across “wild” PDFs, especially multi-page tables.

Your mention of agentic OCR correction and semantic chunking really caught my attention. I’m curious — how did you architect those to stay consistent across diverse layouts without relying on massive rule sets?

Thanks MIke will take a look at Lumen

d1vyank ,

can we connect offline and discuss

thanks Arvinf 94230096(five)(four)

yeah WIP

thanks Ethan will try and embed your feedback

Simple document modelling and classification platform.

no ads in any form

I agree, the browser shouldn't contain ads, but an adblocker is also a basically essential extension that many of us use.

Myself included.

It's basically impossible not to these days.

Nowadays I refuse to fix anyone's computer without installing ad blockers in all installed browsers.

I'd consider it professionally irresonsible to let a non-technical user loose on the web without a good ad blocker. It's just too dangerous out there.

And then you're left with advertorials which are harder to detect.

Java script the good parts is worth a read