Feature Request: AI Text Humanizer for PDFs #22

Motivation

Users often have AI-generated content (essays, reports, articles) saved as PDFs that reads as robotic, repetitive, or obviously AI-written. They need a fast way to rewrite this text into natural, human-sounding language without manually retyping the entire document. A PDF humanizer would save significant editing time and improve content authenticity.

Acceptance Criteria

• User can upload a PDF (single file)
• System extracts readable text from the PDF (handle multi-page docs)
• AI-generated text is rewritten in natural, human-like tone (<15s for typical docs)
• Output preserves original structure (headings, paragraphs, lists)
• Option to adjust humanization strength (light / moderate / heavy rewrite)
• User can download the humanized version as a new PDF (or DOCX)
• Error handling for: password-protected PDFs, image-only scanned PDFs, oversized files

Technical Notes

• PDF text extraction: consider pdfplumber, PyMuPDF, or similar
• Humanization: LLM prompt-based rewriting (tone/style transfer), may require chunking for large documents exceeding context window limits
• PDF regeneration: reportlab or fpdf2 to rebuild formatted output, or export to DOCX via python-docx
• Frontend: display side-by-side original vs. humanized text in a structured panel alongside the upload UI
  Priority

enhancement ai pdf humanizer