Installation¶

From PyPI¶

pip install obsidian-import

This installs the default document backend, anydoc, which handles Word, PowerPoint, Excel, OpenDocument, RTF, EPUB, and CSV input (PDFs stay on the native backend by default). It ships as a compiled wheel, so there is nothing further to install for those formats.

Optional backends¶

Install extras for additional backend support:

# markitdown — fallback for HTML, CSV, and other formats
pip install obsidian-import[markitdown]

# docling — high-quality ML-based extraction for complex layouts
pip install obsidian-import[docling]

Development Setup¶

Clone the repository and install with pixi:

git clone https://github.com/neuralsignal/obsidian-import.git
cd obsidian-import
pixi install

Pre-commit hooks¶

Install pre-commit hooks for automatic linting and formatting:

pixi run pre-commit-install

Running tests¶

pixi run test

Linting and formatting¶

pixi run lint
pixi run format-check
pixi run format

Requirements¶

Python >= 3.12
Core dependencies are installed automatically: pyyaml, click, pdfplumber, pypdf, defusedxml, python-pptx, openpyxl