Converting other formats into PDF
When you convert a Word document to PDF, the converter walks through the document's structure (paragraphs, headings, tables, images) and re-emits each element as PDF drawing instructions on a fixed page.
Image conversions are simpler: each image is wrapped on a PDF page with the correct dimensions, optionally scaled to fit the paper size.
Converting PDF back into editable formats
Going the other direction is harder. A PDF does not store paragraphs or tables, only positioned text fragments. Converters have to group nearby fragments into lines, lines into paragraphs, and detect tables and lists.
For scanned PDFs the text does not exist as text at all, so optical character recognition (OCR) is needed before the file can be turned into editable Word content.
Why output is not always pixel perfect
Word and PDF describe pages differently. Word reflows content based on margins and styles, while PDF places every glyph at a fixed coordinate. Round-tripping between the two requires educated guesses, which is why small layout shifts are normal.
Client-side vs server-side conversion
Modern browsers can run conversion entirely on your device using libraries such as pdf-lib, mammoth, and pdfjs. That keeps your files private because nothing is uploaded.
Server-side conversion can offer higher fidelity for complex documents, but at the cost of sending your file to a third party.