Open-source software we use
Osmanth Docs is built on open-source document-understanding software. We are grateful to its authors.
| Component | Used for | License |
|---|---|---|
| Docling (docling-project) | Layout analysis, reading order, tables, finding formula regions | MIT |
| Docling models: layout (Heron), TableFormer | Model weights | Apache-2.0 / CDLA-Permissive-2.0 |
| PP-FormulaNet plus-M (PaddlePaddle) | Reading equations as LaTeX | Apache-2.0 |
| PDFium / pypdfium2 | PDF parsing, page detection, image extraction | BSD-3-Clause / Apache-2.0 |
| openpyxl | Excel files | MIT |
| KaTeX, marked, DOMPurify, React | Previews in your browser | MIT / Apache-2.0 / MPL-2.0 |