Turn PDFs and documents into ebook source material

Import documents as evidence, not as layout

When you upload a document, Scribora extracts its usable content into the project's source layer. The goal is to understand and reuse the information, not to preserve the visual layout of the original file.

That distinction matters for reports and slide decks: headings, arguments, examples, and facts can inform the new book even though the exported ebook receives its own structure and formatting.

Use documents selectively

• Prefer primary reports and material you are allowed to use.
• Remove obsolete versions when a newer document supersedes them.
• Add your own notes alongside reference documents so the project retains your interpretation and point of view.
• Check quotations, tables, and externally sourced facts before publication.

What the current document parser actually reads

Upload File currently supports PDF, PPTX and DOCX for document ingestion, plus transcripts as TXT, Markdown, VTT or SRT. PDFs are parsed for extractable text; PPTX files are read slide by slide from text frames; DOCX files paragraph by paragraph; VTT and SRT transcripts keep what was said (and who said it, when the VTT names the speaker) without cue numbers or timestamps. Text and subtitle files must be saved as UTF-8 (or as Windows Notepad's "Unicode"); a file in an older encoding is refused with instructions rather than read with mangled characters. This path does not preserve the original visual layout and does not perform OCR on image-only PDF pages.

A file that yields less than 20 characters of extracted text is rejected as unusable. Extracted source text is capped at 100,000 characters before it is stored as a gathered source.