pdf extension support

[furqon]

actually not an issue, but request.
is there any possibilities to add more support on pdf?

[hydropix]

Hi! I’ve been thinking about this. Full PDF-to-PDF translation with exact layout preservation is indeed very hard — I’ve tried and the reconstruction quality wasn’t good enough to ship.
However, I’m considering a "text-only PDF mode" as a pragmatic first step:

Extract clean text from the PDF

Translate it through the existing pipeline

Output as Markdown or plain text (with basic structure: paragraphs, headers)
This would cover books, academic papers, and reports where the content matters more than the exact formatting. Images, complex tables, and precise layout would not be preserved.
I’m away until Monday and won’t be coding before then, but I wanted to float this approach. Would this solve your use case, or do you specifically need a formatted PDF output?

[furqon]

thank you for your consideration, that would be enough. content is matter more than formatting.
output to markdown or docx is enough if possible.