pdf2pdfa

Convert ordinary PDF documents into fully compliant PDF/A files (1b, 2b, 3b).

Installation

pip install pdf2pdfa

Requires Python 3.9+.

CLI Usage

Single file

pdf2pdfa convert input.pdf output.pdf

Choose PDF/A level

pdf2pdfa convert input.pdf output.pdf --level 2b
pdf2pdfa batch *.pdf --level 3b

Supported levels: 1b (default), 2b, 3b.

Batch conversion

pdf2pdfa batch *.pdf

Options

Python API

from pdf2pdfa import Converter

conv = Converter()                    # PDF/A-1b (default)
conv.convert("input.pdf", "output.pdf")

conv = Converter(level="2b")          # PDF/A-2b
conv.convert("input.pdf", "output_2b.pdf")

What it does

• Smart font matching: resolves each PDF font to the best system substitute by family (serif/sans/mono), weight (bold/normal), and style (italic/roman)
• Embeds missing fonts with correct WinAnsiEncoding width metrics
• Attaches sRGB ICC profile with proper /N on the stream dictionary
• Replaces DeviceRGB/DeviceCMYK color spaces with ICC-based equivalents
• Sets PDF/A conformance in XMP metadata (1b, 2b, or 3b)
• Synchronizes XMP and document info dictionary

v3.1.0 Highlights

• New: Multi-level PDF/A support — --level 1b (default), --level 2b, --level 3b
• Python API: Converter(level="2b")
• OutputIntent /S correctly uses /GTS_PDFA1 for all PDF/A levels per ISO 19005

v3.0.0 Highlights

• New: Font matching system — each unembedded font is resolved individually instead of using a single fallback
  • Times-Roman → times.ttf (serif), Courier → cour.ttf (mono), Helvetica → arial.ttf (sans)
  • Bold, italic, and bold-italic variants are matched to the correct system font files
  • Graceful degradation: if an exact match isn't found, falls back through style → weight → category
  • Cross-platform support: Windows, macOS, and Linux font paths
  • --font flag still works as a user override for all fonts
• Refactored: converter.py no longer contains platform-specific font search logic

v2.0.0 Highlights

• Fixed: ICC profile /N entry now correctly placed on stream dictionary (verapdf validation pass)
• Fixed: Font glyph width mismatch for WinAnsiEncoding codes 128-159
• Fixed: DeviceCMYK images now properly covered by CMYK OutputIntent
• New: batch command for multi-file conversion
• New: --validate flag for post-conversion verapdf check
• New: --font flag for custom font embedding
• Removed: reportlab dependency (no longer needed)
• Removed: Python 3.7/3.8 support (minimum 3.9)

Development

git clone https://github.com/nks1990/pdf2pdfa.git
cd pdf2pdfa
pip install -e .[test]
pytest -v

License

MIT - see LICENSE.