BREAKING
HURIDOCS Ships PDF Layout Microservice
0
+
OCR languages
0
+
API endpoints
0
GB
GPU memory
From Messy PDF to Structure
1
Segment elements
↓
2
Detect reading order
↓
3
Export JSON/MD/HTML
↓
4
Local translation
Two Analysis Backends
VGT (default)
Accuracy
●
Vision Grid Transformer
●
Higher precision layout
fast=true
Speed
●
LightGBM classifier
●
Quicker throughput
Ollama Keeps Translation Local
Self-Hosted, Privacy-Conscious
AI NEWS BLITZ
HURIDOCS has released an open-source Docker microservice for PDF layout analysis.