pdf2zugferd

Turn an invoice PDF exported from LibreOffice, OpenOffice or MS Office into a ZUGFeRD / Factur-X (EN 16931) PDF: the original pages plus an embedded, schema-validated factur-x.xml.

The invoice data is read straight from the PDF text layer, so no OCR is needed for office-generated PDFs. Seller data and IBAN come from a config file.

Heads-up: template-based tool. This is not a generic PDF-to-e-invoice converter. It only understands one fixed German invoice layout (see Supported invoice layout) and issues invoices without VAT (Kleinunternehmer, § 19 UStG). PDFs with a different layout are rejected, not guessed. The output is provided as is, without warranty: validate it with your own validator before relying on it for legal purposes.

Install

The tool is published on PyPI as pdf2zugferd (Python >= 3.12). The package, the import name and the command are all called pdf2zugferd.

uvx pdf2zugferd --help          # run without installing
pipx install pdf2zugferd        # or: pip install pdf2zugferd

Usage

cp config.example.toml config.toml   # fill in your data
pdf2zugferd invoice.pdf -c config.toml        # -> invoice_zugferd.pdf
pdf2zugferd invoice.pdf -c config.toml -o out.pdf --xml-only --force

Without Python, download a single-file binary from the GitHub release (Linux, Windows, macOS), or run from source with uv:

uv sync
uv run pdf2zugferd invoice.pdf -c config.toml

Supported invoice layout

The parser targets one fixed template (German labels Rechnungsdatum, Rechnungsnummer, Kundennummer, a position table, Betrag … €, Zahlungsziel beträgt N Tage). Totals are cross-checked and the tool aborts instead of emitting a wrong e-invoice. Invoices are issued without VAT (Kleinunternehmer, § 19 UStG, tax category E).

Limitations

• The output is not converted to PDF/A-3 (fonts are not re-embedded). Most receiving software only needs the embedded XML; strict PDF/A validators may complain. Export the source as PDF/A-1b from LibreOffice if you need it.
• Scanned PDFs are rejected; run them through OCR (e.g. ocrmypdf) first.

Development

uv run ruff format . && uv run ruff check .
uv run pytest                    # unit + e2e (CLI)
uv run bandit -r src && uv run pip-audit --skip-editable

License

Apache-2.0, see LICENSE.