Skip to content

v0.4.0

@Mihailorama Mihailorama tagged this 12 Feb 09:02
Add Nougat and Surya engine adapters (15 engines total).

- Nougat (Meta): academic PDF to Markdown with LaTeX formula support
- Surya: multilingual OCR + layout analysis with bounding boxes, confidence
  scores, and table structure for PDFs and images
- Both engines added to router priority chains (PDF, images, fallback)
- New optional dependencies: docfold[nougat] and docfold[surya]
- Remove GOT-OCR references (no adapter, not well-known)
- 175 tests passing
Assets 2
Loading