📄
AI PDF Metadata Remover
Clear the author, title, creator software and XMP metadata of a PDF while leaving the pages untouched.
Processed locally in your browser · files never uploaded
Drop a PDF, or click to choose
.pdf documents supported, batch OK
Files are processed only in your browser — never uploaded
Up to 10 files · 50 MB each
What metadata removed
All of these fields are cleared — page content, fonts and images stay exactly as they are.
| Field | What it holds | The awkward reveal |
|---|---|---|
Title | The document title, often auto-filled from the first heading or original filename | "resume_v7_FINAL_lowballversion" survives every rename you did in Finder |
Author | The name from the OS account or app license that created the file | Your real name on an "anonymous" submission — or a colleague's name on the proposal you claimed to write |
Subject / Keywords | Description and tags, sometimes inherited from a template | Keywords from a template's original owner, describing a completely different client |
Creator | The application the document was authored in | "Microsoft Word" under a resume claiming years of Adobe expertise |
Producer | The engine that generated the PDF, with version string | A cracked or ancient tool's fingerprint on a professional invoice |
CreationDate / ModDate | When the file was created and last modified, with timezone | A "report we prepared last month" created at 2 a.m. the night before — or dated after the deadline it claims to precede |
| XMP stream | XML metadata: author, history, document IDs, editing software | A second full copy of everything above, which many "cleaners" forget to touch |
PieceInfo | Application-private data attached by authoring tools | Traces of the exact program and workflow that built the file |
Why remove AI metadata?
- Resumes and cover letters. Author and Producer fields shouldn't contradict your story.
- Freelance invoices and contracts. Client documents built on another client's template.
- Blind-review submissions. Your name in the Author field defeats the whole point.
- Legal and dispute documents. Creation dates that contradict a claim are a gift to the other side.
- Anything AI-drafted. Producer strings can reveal the tool that generated the document.
Names inside the document text are a separate problem — read the PDF metadata guide for the full checklist.
Frequently asked questions
- Is my PDF uploaded?
- No. The PDF engine loads on demand and runs entirely in your browser.
- Is the document content changed?
- No — only the Info dictionary, XMP stream and application data are cleared. Pages, fonts and images are untouched.
- Which limits?
- Up to 20 files, 30 MB each. Password-protected PDFs can't be processed — remove the password first.
- Does it remove names written in the text?
- No. Metadata is data about the file; text on the pages stays. Search the document for names before sharing.
- What about tracked changes and comments?
- Those are content, not metadata — accept or remove them in your editor before exporting to PDF.