PDF converter

PDF is a fixed-layout document format: it describes pages, coordinates, fonts, images, annotations, and optional structure, not a simple stream of paragraphs. That is why PDF is excellent for contracts, invoices, forms, proofs, manuals, and anything that must print or archive with predictable page boundaries. Conversion starts by asking what kind of PDF you have. A digitally born PDF usually contains selectable text, embedded fonts, vector drawings, hyperlinks, and bookmarks. A scanned PDF may be only page images with no usable text layer. A hybrid PDF has page images plus OCR text, which can be wrong even when it looks selectable.

Versions and profiles matter. PDF 1.4 through 1.7 are common in office and publishing workflows; PDF 2.0 adds newer semantics that many legacy tools still simplify. PDF/A is an archival profile that restricts external dependencies and expects fonts, color information, and metadata to be self-contained. PDF/X targets print exchange, while tagged PDF is designed for accessibility and reflow. Password protection is separate from profile: some PDFs allow opening but block copying, editing, or printing. Converters should preserve visible pages first, then report which interactive or semantic features could not map to the target.

Readers and writers include browsers, Adobe Acrobat, Preview, LibreOffice, Microsoft Office exports, LaTeX, Typst, design tools, scanners, e-signature systems, and reporting libraries. Those producers make very different PDFs. Office exports tend to convert back to DOCX better than scanned forms. Design PDFs may use absolutely positioned text fragments, glyph subsets, and custom encodings that look correct but extract in the wrong order. Report PDFs often contain table-like text with no actual table model. Forms can be AcroForm fields, XFA packets, flattened appearances, or a mix.

The big conversion gotcha is that PDF is usually an output format, not a source format. PDF to DOCX, Markdown, CSV, or JSON requires reconstruction: reading order, tables, headings, and whitespace must be inferred from page geometry or OCR. PDF to image is predictable because it rasterizes pages. PDF to PDF transformations are safer when they only compress, merge, split, or normalize. When accuracy matters, check whether text is selectable, count pages before and after, inspect tables, and keep the original PDF as the audit copy even when another format is easier to edit.

Plan PDF conversions around the job's evidence requirements. For a court filing, signed agreement, or invoice archive, the PDF itself is usually the record and converted output is a working copy. For a knowledge-base migration, the desired output may be Markdown with headings and links, so small layout differences are acceptable but reading order is not. For accounting, table extraction should be reconciled against totals rather than trusted visually. For accessibility remediation, conversion cannot invent missing tags perfectly; it can only use text, geometry, OCR, and landmarks that exist or can be inferred. Write down those expectations before batch conversion, and store warnings with the output so reviewers know which pages required OCR, reconstruction, or flattening. That note also helps support teams explain why two visually similar PDFs produced different editable files.

Official PDF specification