PDF Metadata Remover

How do I remove PDF metadata and verify the result?

Choose one PDF, inspect the loaded fields, and click Remove All followed by Save Changes. Review Verification of the saved file before clicking Download. This empties the six editable text fields and strips detected document XMP. For additional cleanup, select the checkboxes shown under Also found in this PDF; they start unchecked.

For a worked example, suppose a report displays Title as “Budget draft”, Author as “Alex Morgan”, and Keywords as “internal, budget”. The scan also finds one custom Info key named Department, two embedded files, and three annotation author entries. These are separate groups: clearing Author alone does not clear the others.

Cleanup choices for the example report

Action	Result after saving
Clear Author only	Author becomes empty; Title and Keywords remain editable values.
Remove All	All six text inputs become empty; detected document XMP is stripped on save.
Select all three extra checkboxes	Department is deleted, the embedded-files entry is removed, and the three annotation author entries are cleared.

The extra labels include the detected counts: Remove custom Info keys (1), Remove embedded files (2), and Remove annotation authors (3) in this example. Combine these choices with Remove All when appropriate. The resulting structural scan should find no standard text values, custom keys, document XMP reference, embedded-files tree, or annotation author entries. Review the actual report rather than assuming the button name covers everything.

What is PDF metadata, and which fields can I edit?

PDF metadata describes the document rather than its visible page text. The Info dictionary holds properties such as title and author; document XMP is a separate XML metadata record. PDF structure is defined by ISO 32000, while XMP is standardized in ISO 16684-1. A title written on a page is separate from the Title property edited here.

The six text inputs are Title, Author, Subject, Keywords, Creator Application, and PDF Producer. The last two describe software associated with creating the document. Keywords accepts 1,000 characters; each other input accepts 255. Creation and modification dates appear separately as read-only values. They have no individual edit or clear control.

Changes stay staged until saving. A cleared value shows “Will be removed”; a replacement shows “Modified”. Reset restores the loaded text values, while Use Another PDF leaves the current editing session. If all six loaded text fields are empty, the editor shows an already-clean message instead of inputs. Still inspect any XMP or extra findings: that message concerns editable fields, not every possible identifier.

Why does XMP need separate cleanup?

XMP can contain another representation of authorship, dates, software names, or document identifiers. Clearing an Info field alone would not address that parallel record. This tool removes the detected document-level Metadata reference during saving, including when you retain or replace a Title value. XMP has an inspection panel, not an editor.

The preview extracts simple text elements from five namespace prefixes: dc, xmp, xmpMM, pdf, and pdfaid. It is a best-effort preview, not a complete XML inventory. Nested structures or unsupported representations may produce no displayed fields even when an XMP stream is detected. The interface explicitly reports that situation. An empty preview is therefore different from an absent XMP stream.

What does the JSON snapshot actually contain?

Click Export JSON before changing values to preserve a snapshot of the loaded metadata. Exporting afterward captures your current text edits alongside the dates and XMP fields read during loading. The filename uses the input name followed by -metadata.json. Empty text fields are omitted rather than exported as blank strings.

For example, with only Title and Author populated and no available dates or parsed XMP fields, the export is:

{
  "Title": "Budget draft",
  "Author": "Alex Morgan"
}

There is no enclosing “info” object. Dates, when available, use Creation Date and Modification Date labels with localized English US date-and-time strings. Parsed XMP values sit under an uppercase XMP key, grouped by readable namespace names. Keep this snapshot private too: it deliberately preserves the identifying values you may be removing from the PDF. It is a reference record, not an importable backup.

What do Pass and Check mean after saving?

The saved-file report reopens the exact output bytes. Its seven rows inspect Info fields, custom Info keys, XMP stream, embedded files, annotation authors, incremental updates, and trailer ID. An intentionally retained Title triggers Check for Info fields; that does not mean your edit failed. The Info check excludes dates and Trapped, so passing it does not establish that timestamps were removed.

The incremental-update row counts literal %%EOF markers. More than one produces a warning that older revisions may be recoverable; at most one passes. This marker count is a useful signal, not a forensic guarantee. Likewise, the trailer ID row reports a file identifier; cleanup does not offer a control to delete it. Neither row proves that every identifying byte has been erased.

How is batch cleaning different from editing one PDF?

Selecting at least two PDFs starts sequential cleanup immediately. Batch mode enables every cleanup option, including clearing the entire Info dictionary and its dates. It offers status rows rather than per-file editing or JSON export. A file that fails displays Error while the remaining files continue. Successfully processed files are packaged with Download ZIP; report.pdf becomes report-clean.pdf inside cleaned-pdfs.zip.

Choose single-file mode when retaining selected properties matters. Choose batch mode when the broader cleanup matches your goal, and review any “Cleaned, review needed” status before distributing results. A trailer ID alone does not produce that batch review status.

What does this tool leave outside its cleanup scope?

Processing reads and rewrites PDF bytes in your browser and creates a separate download. It does not redact names printed on pages, inspect visible signatures, or clear comment text. Annotation cleanup deletes author entries, not the annotations themselves. Embedded-file cleanup removes the catalog’s EmbeddedFiles entry; it is not an inspection of every possible associated-file reference or a certified secure-erasure procedure.

Use Redact PDF for sensitive page content or Delete PDF Pages to remove unwanted pages, then inspect the final document again. The upload area advertises PDFs up to 100 MB, but processing still requires browser memory and a readable file. There is no password-entry workflow, so successfully displaying some properties of an encrypted PDF does not guarantee that saving it will succeed.