PDF Glossary

PDF (Portable Document Format) is a file format developed by Adobe that captures a document's text, fonts, images, and layout so it looks identical on any device or operating system. A PDF is designed to be self-contained — everything needed to render the page is embedded inside the file.

PDF/A is an ISO-standardised subset of PDF designed for long-term archival. It forbids features that may not render correctly in the future — JavaScript, audio, video, external font references, and encryption — and requires all fonts and colour profiles to be embedded so the document is self-describing decades later.

PDF/X is a family of ISO standards for graphic exchange in professional printing. It enforces rules around colour spaces (CMYK or spot colours), bleed, trapping, and font embedding so a print shop can send the file straight to press without surprises.

A linearized PDF is reorganised so the first page can be displayed before the entire file downloads. The document streams page-by-page, letting a reader start reading a large file over a slow connection almost immediately.

Metadata is information embedded inside a PDF that describes the document — title, author, subject, keywords, creation date, and the application that produced it. It travels with the file and is read by search engines, document management systems, and viewers.

A PDF defines several rectangular boundaries on each page. The media box is the full physical page size. The crop box is what viewers display. The bleed box extends past the trim for print overflow. The trim box is the final cut size. Together they control what is shown and printed.

Bates numbering stamps a unique, sequential identifier — typically a combination of letters and zero-padded digits — onto every page of a document or set of documents. It is the legal profession's standard way to track and reference pages during discovery and litigation.

Page extraction copies one or more specific pages out of a PDF into a new, smaller file. Unlike deleting, the original document is left intact; the extracted pages form a separate PDF.

Splitting divides a single PDF into multiple output files based on a fixed page count, a page range, bookmark level, or file-size target. Each resulting file is a standalone PDF.

Merging combines two or more PDFs into a single file, appending pages in a chosen order. The original files are not modified; the output is a new combined document.

A watermark is text or an image overlaid on top of — or behind — a page, usually to identify ownership, mark status, or deter unauthorised copying. Watermarks can be faint and centred, tiled across the page, or placed in a corner.

Redaction permanently removes sensitive content — text, images, or regions — from a PDF and replaces it with a black box, ensuring the underlying data cannot be recovered. True redaction deletes the content, unlike merely drawing a black rectangle over it.

Rotation changes the display orientation of one or more pages — typically by 90, 180, or 270 degrees — without altering the page content. It fixes pages that were scanned or exported sideways or upside down.

Cropping trims the visible area of a page by defining a new boundary (crop box), hiding margins, headers, footers, or scanner edges. The content outside the crop box is hidden from view but may still exist inside the file.

Flattening merges all layers — form fields, annotations, comments, and interactive elements — into a single static layer of page content. After flattening, fields can no longer be edited or filled; the entered values become permanent page content.

Reordering lets you drag pages into a new sequence — moving, duplicating, or deleting pages by thumbnail — without re-exporting the source document. It is the visual way to restructure a PDF's page flow.

Encryption scrambles a PDF's contents so only someone with the correct password can open it. Modern tools use AES-256 encryption, the same standard used for classified government data, making brute-force recovery effectively impossible.

A user password (open password) is required to open and view the document. An owner password (permissions password) controls what a reader can do — print, copy text, edit, or comment — even after the document is opened. A document can have one, both, or neither.

A digital signature uses a cryptographic certificate to prove who signed a document and that it has not been altered since signing. Unlike a scanned image of a handwritten signature, a digital signature is mathematically verifiable and tamper-evident.

Certificate-based encryption secures a PDF using a recipient's public-key certificate rather than a shared password. Only the holder of the corresponding private key can decrypt the file, so you can encrypt for specific recipients without transmitting a password.

AcroForm (Acrobat Form) is the original PDF interactive form technology. Fields — text boxes, checkboxes, radio buttons, dropdowns — are defined at fixed positions on the page, and each field has a name and a value that a user fills in.

XFA (XML Forms Architecture) is Adobe's newer, more powerful form technology that supports dynamic layouts — fields that grow, repeat, or reflow based on entered data. XFA forms require Adobe Acrobat and are not fully supported by most other readers.

A form field is an interactive area on a PDF where a user enters data. Common types are text fields, checkboxes, radio buttons, list boxes, combo boxes (dropdowns), and signature fields. Each field has a name, a type, and properties like required or read-only.

Form flattening converts filled-in form fields into static page content, removing the interactive fields and 'baking' the entered values onto the page. After flattening, the data cannot be edited as form fields — it is permanent text on the page.

A bookmark — also called an outline entry — is a navigational link in a PDF's sidebar that jumps to a specific page or view. Bookmarks form a tree, letting readers jump to chapters, sections, or annexes without scrolling.

A PDF can embed other files — documents, images, spreadsheets, even other PDFs — as attachments that travel inside it. A reader can open or save the attached file from within the PDF.

OCR analyses the pixels of an image or scanned page and converts the shapes of letters into searchable, selectable, machine-readable text. Without OCR, a scanned PDF is just a picture of words; with OCR, the text can be searched, copied, and read by screen readers.

A searchable PDF is a scanned document that has been processed with OCR so that a text layer sits invisibly over the page image. The page looks like an image, but the text is selectable and searchable.

A tagged PDF contains structural metadata — headings, paragraphs, lists, tables, alt text — that defines the reading order and meaning of the content. Tagging is essential for accessibility, letting screen readers navigate the document logically rather than guessing from visual layout.

PDF/UA is an ISO standard that defines how to make a PDF fully accessible. It requires tagged structure, a defined reading order, alt text on images, proper table markup, and a document language so assistive technology can interpret the content correctly.

Alt (alternative) text is a short text description attached to an image inside a tagged PDF so that screen readers can convey the image's meaning to users who cannot see it. Good alt text describes the image's purpose, not just what it looks like.

A raster image is made of pixels (like a photo); a vector graphic is made of mathematical paths (like a logo or diagram). PDFs can contain both. Converting a vector PDF page to an image rasterises it — turning crisp, scalable paths into a fixed grid of pixels.

Converting a PDF to an editable Word (.docx) document extracts the text and layout so it can be edited in a word processor. Conversion quality depends on whether the PDF has a real text layer (good) or is a scanned image (requires OCR first).

Converting a PDF to Excel extracts tabular data — rows, columns, and numbers — into a spreadsheet so it can be calculated and manipulated. Tables with clear borders convert accurately; tables with merged cells or no borders require cleanup.

Converting a Word, PowerPoint, or Excel file to PDF produces a fixed-layout, shareable version of the document. The conversion embeds fonts and freezes the layout so it renders identically everywhere, regardless of whether the recipient has the source application.

Converting JPG, PNG, or other image files into a PDF wraps them in a document, optionally one image per page. It is the simplest way to combine photos or scans into a single shareable, printable file.

A colour space defines how colours are represented. RGB (red-green-blue) is for screens; CMYK (cyan-magenta-yellow-black) is for print; grayscale uses shades of grey. Converting between spaces — especially RGB to CMYK for print — can shift colours because the two gamuts don't overlap perfectly.

Lossless compression shrinks a file without discarding any data — the reconstructed file is bit-for-bit identical. Lossy compression achieves much smaller sizes by discarding visual detail the eye barely notices, primarily by downsampling and re-encoding images.

Downsampling reduces the resolution of embedded images — for example from 600 DPI to 150 DPI — to shrink file size. Because screen and standard print rarely need more than 150–200 DPI, downsampling typically has no visible effect while dramatically cutting size.

Font subsetting embeds only the characters actually used in the document rather than the entire font file. Since most documents use a small fraction of a font's character set, subsetting can dramatically reduce file size.

Object streams group many small internal objects (the building blocks of a PDF) into compressed containers, reducing overhead and file size. PDF 1.5+ introduced object streams; tools that use them produce significantly smaller files for the same content.

DPI (dots per inch) measures how many pixels of an image fit in one inch. Higher DPI means more detail but larger file size. PDFs contain images at various DPIs; downsampling high-DPI images is the most effective way to reduce file size.

Document parsing is the automated extraction of structured information — text, tables, key-value pairs, headings — from a PDF so that software can understand and act on its contents rather than treating it as a flat image.

When an AI processes a long PDF, semantic chunking splits the text into meaningful units — paragraphs, sections, or topics — rather than arbitrary fixed-length pieces. This preserves context so the AI can answer questions accurately about specific parts of the document.

An embedding is a numerical representation of a piece of text that captures its meaning. By converting each chunk of a PDF into an embedding, an AI can find the most semantically relevant passages for a user's question — even when the words don't match exactly.

In the context of AI and PDFs, a hallucination is when a language model generates information that sounds plausible but is not actually in the document. Quality AI tools reduce hallucination by grounding answers strictly in retrieved document chunks and citing the source page.

Smart compare analyses two versions of a PDF — typically a contract or report — and identifies every insertion, deletion, and modification, then produces a redlined output with additions highlighted and deletions struck through. AI explains the impact of each change in plain language.

Contract risk scanning uses AI to read a legal document — an NDA, lease, or employment contract — and flag risky clauses, missing standard terms, and obligations that favour one party. It produces plain-language explanations so non-lawyers can spot issues before signing.

Vectorization is the automated detection of table structure — rows, columns, headers, and merged cells — within a PDF so the data can be exported to a spreadsheet or database. AI-assisted vectorization handles tables without visible borders that traditional parsers miss.