pdf-document-translation

This tool translates full PDF documents to Simplified Chinese by converting them into editable HTML intermediates, translating each layer, and rebuilding a final PDF that preserves the original layout, figures, and tables.

Terminology binding in PDF translation generates a document-specific terms.md file before translating any layer, forcing later stages to use consistent terminology throughout the final output for accurate and professional results.

You can translate scanned PDFs because the workflow supports text-based, scanned, and mixed PDFs, automatically choosing the appropriate processing route via document analysis before starting the HTML rebuild.

To resume a failed PDF translation pipeline without starting over, each stage writes checkpointed intermediates under the work directory, allowing you to re-run only the specific failed stage instead of restarting the entire workflow.

The PDF translation workflow enforces a binary QA gate requiring all 16 objective pass/fail checks (G01–G16) to pass before delivery, refusing output and providing clear failure routing if any check fails.

The best way to handle complex PDF layout translation is using a deterministic multi-stage workflow that leverages pdfplumber and weasyprint to extract content, translate layers, and rebuild HTML into a structurally consistent PDF.