liteparse
Local PDF parser with OCR and layout awareness
An OCR model for reading long PDFs, layout-heavy documents, and handwriting. It is aimed at users with NVIDIA GPUs who need consistent OCR speed over many pages.
The model builds on DeepSeek OCR-style work and uses sliding-window attention to keep VRAM use and speed more stable as documents grow. It runs on CUDA/NVIDIA hardware and can process complex layouts rather than just plain text.
Free ai video creation + OCR full books + private photos + more Github finds
Local PDF parser with OCR and layout awareness
Open world models and datasets for robots and physical AI
Open agent skills for documents, PDFs and office files
Convert PDFs, Office files, audio and YouTube into clean Markdown
Turn any PDF into an interactive, quiz-based tutor
Security scanner for third-party AI skills