See what your files reveal. Sweep it out.
MetaBroom opens JPEG, PNG, WebP, PDF, Word, Excel and PowerPoint files and lists the data hiding inside them: author names, computer names, edit history, GPS positions, software versions, comment authors. Then it writes a clean copy and reads it back to prove the fields are gone. Everything happens in your browser.
Free. No upload, no account, no tracking.
Drop files here, or choose them
JPEG, PNG, WebP, PDF, DOCX, XLSX and PPTX — one file or a whole batch. Pasting from the clipboard works too.
Files are read and rewritten inside this browser tab. Nothing is uploaded: after the page has loaded, MetaBroom makes no network requests at all.
How MetaBroom cleans
Reads every metadata container
EXIF, XMP, IPTC, comment and vendor segments in JPEG; text, time and eXIf chunks in PNG; EXIF and XMP chunks in WebP; the PDF Info dictionary, XMP packet and annotation authors; core.xml, app.xml, custom.xml, comments, tracked changes, people.xml and rsids in Office files.
Removes without re-encoding
Images are rebuilt segment by segment and the pixel data is copied byte for byte, so there is no quality loss. PDFs are rewritten from their parsed objects with the same page content. Office documents keep every content part untouched.
Reads the clean copy back
After cleaning, the new file is parsed a second time and the After card lists exactly what that read found. “Cleaned” is a measurement, not a promise — and if a field survives, you see it in amber.
Works offline, in batches
Drop any number of files and download them as one ZIP. Everything runs in the page: after it has loaded, MetaBroom works in airplane mode and never contacts a server.
About MetaBroom
MetaBroom is a metadata viewer and metadata cleaner that runs entirely in the browser. Drop a photo and it lists the EXIF, XMP and IPTC fields: camera model, serial number, lens, capture time, GPS position, the artist tag and the software that last touched the file. Drop a PDF and it reads the Info dictionary and the XMP packet — title, author, producer string, creation and modification dates, annotation authors — and counts the incremental updates that may still hold earlier versions. Drop a Word, Excel or PowerPoint file and it opens the ZIP container to show core.xml, app.xml and custom.xml: creator, last modified by, company, template name, total editing time, revision number, comment and tracked-change authors, the people list and any embedded thumbnail.
Cleaning is deliberately conservative. JPEG, PNG and WebP files are rewritten segment by segment, so the pixels are copied unchanged and there is no second round of lossy compression — a lossless EXIF remover rather than a converter. PDFs are rebuilt from their parsed objects with blank Info entries and no XMP, which also flushes data left behind by earlier saves. Office documents get fresh docProps parts and lose custom properties and thumbnails, while the text, formatting, sheets and slides stay exactly as they were. Every clean copy is read again before you download it, batches come back as a single ZIP, and nothing is uploaded at any point.
Guides & articles
Where metadata hides in documents and photos, what it reveals, and how to remove it.
What metadata hides in Office documents
A .docx is a ZIP archive with a properties folder inside. What core.xml, app.xml, comments and tracked changes record — and how to clean it.
6 min read/August 2026
PDF metadata explained: Info dictionary and XMP
A PDF keeps metadata in two places, plus dates with time zones, producer strings, images with their own EXIF and versions you thought were gone.
7 min read/June 2026
Metadata leaks that made headlines
A dossier, a fugitive's photo, a redacted report and a printed page: five documented cases where the file said more than its author meant to.
7 min read/May 2026
Frequently asked questions
Are my files uploaded anywhere?
No. MetaBroom reads, rewrites and re-checks every file inside your browser tab using the File API and pure JavaScript parsers. After the page has loaded it makes no network requests at all — you can switch to airplane mode and it keeps working. Nothing is stored either: closing the tab discards everything.
Which file formats does it handle?
JPEG, PNG and WebP images; PDF documents; and the Office Open XML formats DOCX, XLSX and PPTX. The format is detected from the bytes, not the extension, so a renamed file is still recognised. Not supported: HEIC/AVIF photos, camera RAW files, encrypted PDFs and the legacy binary .doc/.xls/.ppt formats — convert those first.
Does cleaning reduce image quality?
No. MetaBroom never decodes or re-encodes pixels. A JPEG is parsed segment by segment and the EXIF, XMP, IPTC, comment and vendor segments are simply left out of the copy; the compressed image data is copied byte for byte, so there is no second generation of JPEG loss. PNG chunks and WebP chunks are handled the same way. The clean file usually opens pixel-identical to the original, only smaller.
What is the ICC profile option for?
An ICC profile tells viewers how to interpret the colours (sRGB, Display P3, Adobe RGB and so on). It rarely contains anything personal, so it is kept by default: removing it can make wide-gamut photos look duller. Switch the option off if you want every non-pixel segment gone; the profile description itself — for example “Display P3” — does hint at the device that produced the file.
Are pictures inside a PDF cleaned too?
No. A PDF stores pictures as image objects, and some converters copy a JPEG into the PDF without touching it, so its own EXIF — including GPS — can survive inside the document. MetaBroom clears the PDF's Info dictionary, XMP packet and application data and rewrites the file, but it does not open the image streams. If a PDF was built from photos, clean the photos first or re-export the PDF from a clean source.
What happens to comments and tracked changes in Word files?
They stay in the document — MetaBroom does not change the text or the review marks. With “Blank comment and revision authors” switched on, the author name, initials and date on every comment, insertion, deletion and formatting change are blanked, the people list is emptied and the revision save IDs (rsids) are removed. If the comments themselves should not be shared, accept or delete them in Word first, then clean the file.
Can I clean many files at once?
Yes. Drop a whole batch, then press “Clean all and download ZIP”. Files are processed one after another with a progress bar, and the clean copies arrive as a single ZIP archive with “-clean” added to each name. Mixed batches are fine — photos, PDFs and Office files together.
How can I verify the metadata is really gone?
MetaBroom does it for you: every clean copy is parsed again before it is offered for download, and the After card lists exactly what the second read found — normally nothing. You can also drop the downloaded file back into the tool, or check it with an independent reader such as ExifTool (exiftool -a -G1 file). Both should show the same empty result.
Why is the cleaned PDF a different size?
Because it is rebuilt rather than patched. The original objects are parsed and written out again with compressed object streams, so a PDF that had grown through incremental saves often shrinks a lot — those appended layers, which can still hold deleted text or earlier titles, are not carried over. Occasionally a small file grows slightly because of the new cross-reference structure. The pages themselves are unchanged.