People who share material with a reporter often think about the story and forget the file. A document, a photo, a recording or a spreadsheet can name its author, its owner, the device, the office network folder and the time it was made. Reporters who receive files have the same problem in reverse: what they publish can expose the source. This guide is practical, not legal advice, and it is not a substitute for proper security training.
What files commonly reveal
- Documents: the author and last editor, the company name from the licence, comments and tracked changes with names, the folder path of the file, and printing dates.
- Photos and video: GPS, the phone or camera model, a serial number, and the exact time.
- PDFs: the program that created them and the author field, sometimes the original file name.
- Audio: the recording app, the device and the date.
- Spreadsheets: hidden sheets, saved-from paths and links to other files on a network drive.
If you are a source
- Work on a copy and keep the original somewhere safe.
- Remove metadata with the matching tool: Document, PDF, Image, Video or Audio.
- Read the report. Content that names people, such as comments and tracked changes, is listed, not deleted, so you can decide.
- Open the cleaned file in the File Metadata Viewer to check what is left.
- Think about the content too. Metadata is one trail; wording, a unique detail or the fact that few people had access are others.
If you are a reporter
Do not open a source’s file on a work computer that is logged into your accounts if you can avoid it, and do not forward it by email. Read its metadata before you decide what to do. If you publish a document, publish a clean copy, and remember that watermarks, fonts and layout quirks can also point to the origin.
A note on how this site works
Documents, images, SVG, GIF, EPUB and ZIP files are cleaned inside your browser, so they are not uploaded. Video, RAW, HEIC and audio need a server and are deleted a few minutes after cleaning. The File Metadata Viewer reads files on the server and deletes them at once. The page for each tool says which applies.