Edit PDF Metadata

Edit PDF Metadata shows and changes a PDF's document properties — the Title, Author, Subject, Keywords, Creator and Producer text plus the Creation and Modified dates. It is for the everyday tidy-up: a report whose title still says "Microsoft Word - draft", a scan with the wrong author, keywords left over from a template, or a file you are about to publish whose Producer advertises the software that made it.

Drop a PDF and its current values appear in the fields, ready to edit. Change what you want and click Apply; anything you leave alone keeps its value. Emptying a field clears that one entry, and ticking "Clear all metadata" empties all six text entries and removes both dates in one go. The edited file downloads as <name>-metadata.pdf, and your original is never changed.

What exactly is edited, and what is not, matters here. A PDF can carry metadata in more than one place, and this tool reads and writes only the standard document information dictionary — the /Info entries described in PDF 32000-1 §14.3.3 (Title, Author, Subject, Keywords, Creator, Producer, CreationDate, ModDate). Many files also carry an XMP metadata packet: an XML description defined by ISO 16684-1 that often repeats the same title and author. XMP is stored in a metadata stream, not in /Info, and this tool does not read, change or remove it. Nor does it touch metadata kept anywhere else in the file. So editing here updates the properties most readers show in their Document Properties box; it is not a way to strip every piece of identifying information from a PDF.

Clearing deserves the same plainness. "Clear all metadata" writes empty values into the /Info entries and removes the two date entries — it does not scrub the old text out of the file's bytes, and a tool that inspects the raw file could still find traces. Some PDF writers also add their own Producer string whenever a document is opened and saved, so an app that re-saves this file may put a Producer back after you cleared it. Treat clearing as an edit that blanks the visible properties, not as a guarantee that the information is gone.

The edit changes only /Info, so page content, page sizes, bookmarks, links and form fields are kept as they were. The whole file is re-saved, though, so a digital signature that covered the original bytes will no longer validate. Password-protected and permission-locked PDFs are refused with a clear message, because the library used here cannot decrypt them; files up to 200 MB are accepted. Everything runs in your browser with pdf-lib — your PDF is not uploaded, and you can confirm in your browser's developer tools (Network tab) that no request carries the file.

1. The file's first bytes are checked for the PDF signature, then pdf-lib opens it with metadata updating switched off, so the library itself does not stamp the file; encrypted and damaged files are refused with a specific message. 2. The eight /Info entries are read with pdf-lib's getters and shown in the fields; an entry that is missing — or stored as something that is not a string — reads as blank rather than failing the read. 3. A date you type is normalised to the PDF date form D:YYYYMMDDHHmmSS in UTC. "2026-01-15" becomes D:20260115000000Z and "2026-01-15 14:30" becomes D:20260115143000Z — the trailing Z is the UTC designator pdf-lib writes; a value typed with an offset is first converted to UTC. Text that is not a real date is refused with a message naming the accepted forms, before the file is touched. 4. Keywords are split on commas and whitespace, trimmed, and de-duplicated without regard to case; the PDF stores them as one space-separated string. 5. Apply writes the fields you changed (or empties them all if Clear all metadata is ticked), re-saves the document and offers <name>-metadata.pdf. Nothing outside the /Info dictionary is altered.