Sanitize PDF.

Each cleanup category shows a count and an expandable list. Metadata entries can show their stored values, external links show their destinations, and ordinary form fields are listed by name. The first 40 details per category are displayed, with long values shortened; cleanup covers the whole category.

Counts describe properties, references, actions and fields. An attachment can have both a file reference and an embedded-data object, so a category count is not necessarily the number of files or threats. Finding a script does not mean it is malicious.

Scripts & actions removes recognized JavaScript, automatic open and event actions, launch and form-submission actions, remote-file actions and external stream references. Removing these features can disable calculations, automatic behavior or interactive navigation that depended on them.

Attachments & media removes embedded files, attachment annotations, multimedia and associated references. Website and email links have their own optional category. Ordinary links and bookmarks pointing to pages inside the document are retained where their destinations remain valid.

Metadata includes the document-information dictionary, custom properties, standard XMP metadata references and the PDF file identifier. Comments & annotations removes notes, highlights, stamps and other markup while keeping form widgets and internal page links. Attachments and external links are handled by their separate categories.

Names or confidential details printed on a page remain in the document. Metadata cleanup does not remove every possible identifier, embedded-image EXIF or data hidden inside ordinary page resources. Removing redaction annotations does not apply redaction. Use Redact PDF when the visible or underlying page content must be permanently removed.

The optional Form answers category resets supported ordinary text, choice, checkbox and radio fields. It removes stored current, default and rich-text answers, and replaces text and choice appearances with empty appearances so old answers are not left visibly on the page. Field names, available choices and their basic formatting remain.

This is different from flattening a form, which keeps answers visible and removes interactivity. Signature fields are not reset as form answers. Dynamic XFA forms are unsupported, and malformed or ambiguous fields can prevent this category from being used. Some field-border styling may change when an empty appearance is rebuilt.

Start creates a full rewritten PDF and removes unreferenced objects. The saved result is reopened to check the selected supported categories, the page count, and cleared form values and rebuilt text or choice appearances. A failed cleanup check does not produce a cleaned download.

The report lists removals or reset fields for selected categories and remaining counts for unselected categories. Related structures can be removed together. The PDF preview is for checking appearance; it does not execute document scripts or open attachments. Back preserves your choices so you can create a different copy from the original.

The operation edits PDF structures rather than converting pages to images. Ordinary text, Arabic text, fonts, images and vector page content remain in their original form. A scanned PDF can be cleaned without OCR. Selected annotations or form answers can change the page appearance.

Existing digital signatures no longer validate the rewritten document. Removing metadata or attachments can also break PDF/A requirements, so do not treat the result as archival validation. Password-protected PDFs and dynamic XFA forms are unsupported. Complex or unusually large structures can reach resource limits before the stated file or page limit.