Appendix E — Glossary
Summary
Plain-language meanings for the words this book uses, in alphabetical order. Where a term is tied to an app, the app is linked. These are the book's working definitions, not legal or standards-body definitions.
Details
Accordion fold — A print layout where the pages keep natural order and the sheet folds zig-zag. One layout of Booklet.
AI (in this book) — A language or image model used by an app. AI results are drafts: they can be wrong or differ between runs. See Trust.
Annotation — A comment, highlight or markup placed on top of a PDF page. Search adds highlights as annotations; Parser extracts them.
API (REST) — The way a program talks to Documentize: submit a task, poll its status, download the result. See REST API.
Archive file — A container such as ZIP, RAR, 7z or TAR that holds other files. Merger and Converter unpack supported ones.
ATS (applicant tracking system) — Software that reads CVs automatically. One mode of Resume Analyzer shows what an ATS would extract.
Base64 — A way to write binary data as plain text. Converter has a box for pasting it.
Bookmark — See Outline.
Booklet — Pages reordered so that, printed double-sided and folded, they read in order. See Saddle-stitch.
Certificate (.pfx) — A file holding your signing key and identity, protected by a password. Needed by Sign.
Citation — In Ask PDF, a marker like [1] linking an answer to a page and a highlighted quote.
CLI (command-line interface) — Running a plugin by typing a command. See CLI plugins.
Continuation — A way to finish a partly processed batch later, free of charge. If a free-plan run stops at its time limit, the leftover files are saved and offered after a cool-down (default 12 hours, kept 2 days). See limits and If a task stops early.
Cool-down — The wait before a continuation can start (default 12 hours; environment-configured).
CSV — A plain-text table with values separated by commas. Output of Extract Metadata, Extract Form Data and Page SEO Checker.
Deskew — Straightening a scan whose text lines are tilted. See Deskew.
Digital signature — A cryptographic stamp showing a file has not changed since signing and was made with a certificate's key. Not the same as a drawn signature image. See Sign.
DPI (dots per inch) — How much detail an image holds per inch. Higher is sharper and bigger. Compressor lowers it to 150, 120 or 96.
Duplex — Printing on both sides of the sheet. Booklet has Auto and Manual duplex.
Embedded font — A font stored inside the PDF so it looks the same everywhere. PDF/A requires it. See PDF/A Validator.
Encryption — Scrambling a file so it needs a password. See Protector and Unlock.
EXIF — Hidden data in photos (camera, date, location). The "Public release" policy of Recovery removes it from images.
Flatten — Turning interactive parts (form fields and their values) into fixed page content. See Flatten.
Form field (AcroForm) — A fillable box, checkbox, radio button or drop-down in a PDF. See Extract Form Data.
Grid / Strip layout — Merger modes that place images in a grid or in vertical/horizontal blocks.
Imposition — Arranging pages on a printer sheet so they come out in the right order after folding or cutting. Booklet and tile mode in Resize are forms of it.
ICC profile / output intent — Colour information PDF/A requires. Missing ones are a common reason for "not valid".
JPEG — A lossy image format. Compressor re-saves pictures as JPEG at quality 75.
Knowledge base (PDF KB) — A small linked wiki built from your PDFs. See PDF Knowledge Base.
Layer (optional content) — A switchable group of drawing content inside a PDF. Protector can lock, merge, extract or flatten layers.
Linearization (Fast Web View) — Arranging a PDF so a browser can show page 1 before the whole file has loaded. The site describes Optimize as doing this.
Lossless / lossy — Lossless keeps content exactly; lossy discards detail to save space. See II.7 Quality model.
MCP (Model Context Protocol) — A standard that lets an AI assistant call tools. The Documentize MCP server exposes document operations. See MCP.
Metadata — Information about a document (title, author, dates, producer). See Extract Metadata.
NuGet — The package system for .NET libraries. See NuGet Updater.
OCR (optical character recognition) — Software reading text from images of text. The result is an estimate. See OCR Image and Searchable.
Operation — One processing call against your daily allowance. In MCP, one operation is one processing call whatever the file count.
Outline / bookmark — The navigable tree in a reader's side panel. Made by Add TOC and TOC AI.
Owner / user password — A PDF can have a password to open it and another for permissions. Protector uses one password for both.
PDF/A — A family of standards for long-term archiving (levels such as 1B, 2B, 3U). See PDF/A Validator.
PDF/UA — The accessibility standard for PDF; the validator can check PDF/UA-1.
PKCS#12 — The file standard behind .pfx certificates.
Plugin — A small command-line program that does one operation. See Plugins.
Pull request (PR) — A proposed change to a repository for a person to review. Several developer apps deliver one.
RAG (retrieval-augmented generation) — An AI answers from passages retrieved from your documents instead of from memory. Ask PDF works this way.
Redaction — Permanently removing sensitive content (not just covering it). Not available as an app; see PDF Redactor.
Regular expression — A pattern language for text search, optional in Search.
Retention — How long results stay on the server: about two days for most apps; about a day for Ask PDF and PDF KB. See limits.
REST — See API.
Saddle-stitch — Binding where folded sheets are nested and stapled at the fold. The booklet layout in Booklet.
Sanitization — Stripping hidden data (history, metadata, attachments) from a file. Recovery offers seven policies.
Scanned PDF (image-only) — A PDF whose pages are pictures with no text. Make it searchable with Searchable.
Self-host — Running the Documentize engine on your own server. See Self-host.
Seam — In Refactor Assistant, a part of a big class that could become its own class.
Sitemap — A file listing a site's pages. Page SEO Checker can start from one.
SVG — A vector image format built from shapes. See SVG Maker.
Text layer — The invisible or visible real text in a PDF. Without it you cannot search, select or copy. Searchable adds one.
Tile mode — Putting several source pages on one sheet. See Resize.
TOC (table of contents) — A contents page and/or outline. Manual in Add TOC, automatic in TOC AI.
UML / .puml — A diagram text format; Assembly Chat writes diagram sources in it.
Unicode — The universal text encoding; PDF/A "U" levels require text to be extractable as Unicode.
Watermark (visible vs hidden) — Visible: stamps like "DRAFT". Hidden: invisible markers in AI-generated text. Unwatermark addresses only the hidden kind.
Wiki language / wiki links — Settings and structure of the knowledge base built by PDF KB.
X-ray report — The detailed repair report from Recovery.
XFDF / XML — Technical formats in which Parser writes annotations and form data.
ZIP — The container in which several results are delivered.