Loading...
Product updates, engineering deep-dives, and opinions on AI for documents.
All articles (5)
Subscribe to updatesA deep dive into our OCR pipeline: layout-aware segmentation, multi-pass decoding, and why we ditched monolithic models for ensembles.
Names in multiple languages, handwritten signatures, medical codes, context-dependent identifiers. Here's the taxonomy that powers Smart Redact.
How hierarchical retrieval, section-aware chunking, and persistent memory let our chat handle book-length documents gracefully.
Chain tools, schedule runs, trigger via webhook. Your document automation, now shareable and version-controlled across the team.
What it took, what changed, and why we took the long route — with AES-256 at rest, TLS 1.3 in transit, and zero data retention for enterprise.