Multi-Format Documents

paperjam is not limited to PDF. It can open, extract content from, and convert between six document formats: PDF, DOCX, XLSX, PPTX, HTML, and EPUB.

Opening any document

paperjam.open() detects the format automatically from the file extension or magic bytes. For PDF files it returns a Document; for everything else it returns an AnyDocument:

import paperjam

# PDF — returns Document
pdf = paperjam.open("report.pdf")

# DOCX — returns AnyDocument
docx = paperjam.open("proposal.docx")

# XLSX — returns AnyDocument
xlsx = paperjam.open("budget.xlsx")

This works with in-memory bytes too. When the format cannot be inferred from a file extension, pass it explicitly:

doc = paperjam.open(raw_bytes, format="docx")

Format detection

If you need to know the format before opening, use detect_format():

fmt = paperjam.detect_format("mystery_file")
print(fmt)  # "pdf", "docx", "xlsx", "pptx", "html", "epub", or None

It inspects magic bytes first, then falls back to the file extension. Returns None when the format is unrecognised.

Document vs AnyDocument

paperjam has two document classes:

Class	Returned for	Capabilities
Document	PDF	Full feature set: forms, signatures, rendering, annotations, redaction, encryption, page manipulation, and everything in AnyDocument
AnyDocument	DOCX, XLSX, PPTX, HTML, EPUB	Text extraction, table extraction, structure extraction, image extraction, Markdown conversion, format conversion

PDF-specific operations like redact_text(), encrypt(), fill_form(), sign(), render_page(), and add_watermark() are only available on Document. The extraction and conversion methods are shared.

Opening each format

PDF

doc = paperjam.open("report.pdf")
print(doc.page_count)
print(doc.pages[0].extract_text())

DOCX (Word)

doc = paperjam.open("proposal.docx")
print(doc.page_count)
print(doc.extract_text())

XLSX (Excel)

Each sheet is treated as a page. Tables are extracted natively from the spreadsheet structure:

doc = paperjam.open("budget.xlsx")
print(f"{doc.page_count} sheets")

tables = doc.extract_tables()
for table in tables:
    print(table.to_csv())

PPTX (PowerPoint)

Each slide is a page:

doc = paperjam.open("deck.pptx")
for i in range(doc.page_count):
    text = doc.extract_text(pages=[i + 1])
    print(f"Slide {i + 1}: {text[:80]}...")

HTML

doc = paperjam.open("article.html")
md = doc.to_markdown()

EPUB

doc = paperjam.open("book.epub")
print(doc.metadata)
text = doc.extract_text()

Common extraction methods

All document types (both Document and AnyDocument) support these extraction methods:

doc = paperjam.open("any_file.docx")

# Plain text
text = doc.extract_text()

# Text with line positions
lines = doc.extract_text_lines()

# Tables
tables = doc.extract_tables()
for table in tables:
    print(table.headers)
    for row in table.rows:
        print(row)

# Structured content (headings, paragraphs, list items, tables)
blocks = doc.extract_structure()

# Embedded images
images = doc.extract_images()

# Markdown conversion
md = doc.to_markdown()

Context manager support

Both Document and AnyDocument support the context manager protocol for automatic cleanup:

with paperjam.open("data.xlsx") as doc:
    tables = doc.extract_tables()
    # resources are released when the block exits

When to use AnyDocument directly

In most cases, paperjam.open() is all you need. If you want to be explicit, you can instantiate AnyDocument directly:

from paperjam import AnyDocument

doc = AnyDocument("report.docx")
doc = AnyDocument(raw_bytes, format="pptx")

See the AnyDocument API reference for the full method list and the Format Conversion guide for converting between formats.