PDF Metadata Editor & Remover

It is a small set of descriptive fields stored alongside the pages rather than on them: title, author, subject, keywords, the software involved, and dates. Most people meet it in three places. A PDF reader shows it in a document properties panel, and many readers put the title — not the file name — in the window or tab heading. Search tools on your computer and document management systems index those fields, so a well-titled file is findable and an untitled one tends not to be. And a PDF published on a website often has its title used as the heading in search results.

Because it was filled in for you. When a PDF is exported from Word, Pages or Google Docs, the author is taken from the account or user name configured in that application — often the person whose template you started from, or whoever set up a shared machine. Scanners write the device or driver name instead. Nothing prompts you at export time, so the same value travels with every file you produce until you change it at the source. Correcting one PDF corrects that PDF; correcting the user name in the application that exports them stops the problem recurring.

No, and this is the mistake worth avoiding. Document properties are one thin layer. A name can also sit in the visible text, in a header or footer, in a signature block, in comments and annotations, in an embedded attachment, and in the file name itself — which is usually the first thing a recipient sees and is not stored inside the PDF at all. Scanned pages can carry device identifiers of their own. If a document has to reach someone who must not know who wrote it, read the pages as a stranger would and rename the file as well.

Nowhere. The PDF is parsed and rewritten by a JavaScript library running inside the page, so the bytes stay in the tab you have open and no copy is sent anywhere to be processed. The edited version is built in memory and handed to you through a temporary link the browser creates for it, which dies with the tab — reload or close the page and both the original and the edited copy are gone, with nothing left to clean up. That matters more than usual here, because the documents whose properties people want to change tend to be contracts, applications and CVs.

A PDF can describe itself twice. The document Info dictionary holds plain title, author, subject and keyword entries, while an XMP packet can repeat them as XML. Nothing forces the two stores to agree, and software differs on which it trusts. Apply & Download changes the Info fields and explicitly tells you when an original XMP packet remains. If you need both stores gone rather than edited, use Remove document info + XMP and download the separately verified copy.

Two causes account for nearly all of it. The file is encrypted — either a password needed to open it, or an owner password restricting changes, both of which stop a library reading the contents until the password is supplied, and there is no box here to supply one. Or the file is damaged: truncated by an interrupted download, or written by software that produced a structure the parser will not accept. The message you get does not separate the two cases. Opening the file in an ordinary PDF reader is the quickest test — if that complains too, it is broken rather than locked.

By convention the Creator names the application the document originally came from — a word processor, design tool or scanner — while the Producer names the component that wrote the PDF bytes, typically a library or export engine. A file drafted in one program and passed through another commonly shows one of each. Both fields are visible and editable here, but changing them only changes labels; it does not erase visible content or prove a different origin.

Not from this page — the creation date is shown read-only and there is no field for it. In the format itself the date is an ordinary entry that any tool with write access can set, which is worth remembering in both directions: it is not proof of when a document was made, and changing it would not hide much, since files usually carry other traces of their history. Note also that both dates here are rendered in your own time zone and locale, so a colleague elsewhere can read a different clock time from the very same file.

Not by hand, past about a dozen. This page handles a single file per pass, which is fine for the occasional fix and tedious for a folder of a hundred. For bulk work the usual answer is a command line tool such as ExifTool or pdftk, driven by a short script that walks the folder and applies the same title, author or keyword set to every file. Many document management systems can also write these fields on ingest, taking the values from the catalogue record rather than the file. Work on copies, because a bulk edit is awkward to undo.