Fair questions, straight answers.

Yes — by architecture, not by policy. The application contains no networking code at all: no HTTP client, no sockets, no auto-updater, no crash reporting. There's no “disable telemetry” setting because there's no telemetry to disable.

To verify: run it with any firewall in block-and-notify mode (Windows Firewall, GlassWire, Portmaster…) — you'll see it never attempts a connection. The source code is also auditable: the scanning engine has zero third-party dependencies.

No. Files are read from your disk, analysed in memory on your PC, and the cleaned copy is written back to the same folder. Nothing leaves your computer — the app has no way to send anything anywhere.

Cleaning never touches the original file — it writes a separate Name (cleaned).ext copy, so there is always a fallback. The cleaned copy preserves all visible content: text, images (pixels untouched), formatting and layout. If a file is damaged or unusual enough that a safe rewrite isn't possible, PDF Doc Shield refuses and tells you, rather than producing a questionable output.

PDF (.pdf) and modern Word documents (.docx, .docm, .dotx). Legacy binary .doc files aren't supported yet — open them in Word and save as .docx first. Excel and PowerPoint support is on the roadmap.

PDF Doc Shield doesn't bypass protection. An encrypted PDF is reported as such and can't be scanned in depth or cleaned. Remove the password in your PDF viewer first, then scan.

Usually the opposite — a blank author field is unremarkable, while your full name in a file you meant to share anonymously is not. That said, everything is selectable: for a resume you might keep the Title; for a public records request you might scrub everything. You choose, finding by finding.

Any element that makes your document contact a server when opened. The classic is a 1×1 invisible image loaded from a remote URL: each time the file opens, the request tells the sender when you read it, how many times, and your network location. Word field codes, remote templates and PDF auto-run actions can all do this — silently. PDF Doc Shield flags them as CRITICAL and removes the remote reference without changing the layout.

Built-in inspectors focus on the obvious property fields. PDF Doc Shield also reads the places they skip: the hidden XMP metadata packet in PDFs, RSID editing-session fingerprints, document GUIDs, relationship targets with local file paths, EXIF/GPS inside embedded images, PDF incremental-update history, and compressed object streams. It parses the whole file, not just the metadata summary.

Cleaned PDFs are rebuilt from scratch: only the current document's object graph is written, so previous revisions and orphaned data aren't “erased” — they're simply never written into the new file. Word documents are repackaged with the offending parts removed and rewritten. As a final check, every cleaned file is automatically re-scanned and must come back clean before the app reports success.

PDF Doc Shield is free to download and use. No account, no registration, no trial timer — those would all require the kind of infrastructure (and data collection) this tool exists to avoid.

Windows 10 (64-bit) or Windows 11. The download is self-contained — no .NET installation or other prerequisites needed. It runs fine without administrator rights.

Scanning is read-only: the file is opened, analysed in memory and closed — nothing is modified, and nothing is transmitted (there's no code capable of transmitting). The optional recent-scans list is stored only on your PC and can be cleared by deleting one local file.