PDF to Word: Convert PDF to Editable .docx (Free, No Upload)

# Install LibreOffice if not already installed
# Ubuntu/Debian: sudo apt-get install libreoffice
# macOS: brew install --cask libreoffice
 
# Convert PDF to DOCX via command line (headless)
libreoffice --headless --convert-to docx input.pdf
 
# Convert to DOCX and place output in specific directory
libreoffice --headless --convert-to docx --outdir ./output input.pdf
 
# Batch convert all PDFs in a folder
libreoffice --headless --convert-to docx *.pdf

from pdfminer.high_level import extract_pages
from pdfminer.layout import LTTextContainer, LTChar, LTAnno, LTFigure
from docx import Document
from docx.shared import Pt, RGBColor
from docx.enum.text import WD_ALIGN_PARAGRAPH
import re
 
def pdf_to_docx(pdf_path: str, docx_path: str):
    doc = Document()
 
    for page_layout in extract_pages(pdf_path):
        page_height = page_layout.height
 
        for element in page_layout:
            if not isinstance(element, LTTextContainer):
                continue
 
            text = element.get_text().strip()
            if not text:
                continue
 
            # Detect heading vs body text by font size
            max_font_size = 0
            for char in element:
                if isinstance(char, LTChar):
                    max_font_size = max(max_font_size, char.size)
 
            if max_font_size >= 18:
                para = doc.add_heading(text, level=1)
            elif max_font_size >= 14:
                para = doc.add_heading(text, level=2)
            else:
                para = doc.add_paragraph(text)
 
    doc.save(docx_path)
    print(f"Saved: {docx_path}")
 
 
# More sophisticated: extract with character-level formatting
def pdf_to_docx_formatted(pdf_path: str, docx_path: str):
    doc = Document()
 
    for page_layout in extract_pages(pdf_path):
        current_para = None
        last_y = None
 
        for element in sorted(page_layout, key=lambda e: -e.y1):
            if not isinstance(element, LTTextContainer):
                continue
 
            text = element.get_text()
            if not text.strip():
                continue
 
            # New paragraph if significant vertical gap
            if last_y is not None and abs(element.y1 - last_y) > 20:
                current_para = None
 
            if current_para is None:
                current_para = doc.add_paragraph()
 
            run = current_para.add_run(text)
            last_y = element.y0
 
    doc.save(docx_path)

pip install pdfminer.six python-docx

# Install Pandoc
# Ubuntu: sudo apt-get install pandoc
# macOS: brew install pandoc
 
# Direct PDF to DOCX (uses pdftotext internally)
pandoc input.pdf -o output.docx
 
# With better formatting via reference document
pandoc input.pdf -o output.docx --reference-doc=template.docx
 
# Extract text from PDF first, then convert
pdftotext -layout input.pdf - | pandoc -f plain -t docx -o output.docx

from pdf2docx import Converter
 
def convert_pdf_to_word(pdf_path: str, docx_path: str, start: int = 0, end: int = None):
    cv = Converter(pdf_path)
    cv.convert(docx_path, start=start, end=end)   # start/end are 0-based page indices
    cv.close()
    print(f"Converted: {pdf_path} → {docx_path}")
 
# Convert all pages
convert_pdf_to_word("report.pdf", "report.docx")
 
# Convert only pages 0–4 (first 5 pages)
convert_pdf_to_word("report.pdf", "report-first5.docx", start=0, end=5)

pip install pdf2docx

from pdf2image import convert_from_path
import pytesseract
from PIL import Image
from docx import Document
 
def scanned_pdf_to_docx(pdf_path: str, docx_path: str, dpi: int = 300, lang: str = "eng"):
    pages = convert_from_path(pdf_path, dpi=dpi)
    doc = Document()
 
    for i, page_img in enumerate(pages, start=1):
        text = pytesseract.image_to_string(page_img, lang=lang)
        
        # Add each OCR'd line as a paragraph
        for line in text.splitlines():
            if line.strip():
                doc.add_paragraph(line)
        
        if i < len(pages):
            doc.add_page_break()
 
    doc.save(docx_path)
    print(f"OCR'd and saved: {docx_path}")