How to see the hidden metadata in any file
Every file you share carries a description of itself that you never see on screen. Here is how to read it, from the viewers already on your computer to a full inspection, and how to tell which findings actually matter.
Metadata is data about a file rather than the content of it: where a photo was taken, which phone took it, who wrote a PDF and with what software, when a video was recorded. Apps write it automatically and almost none of it is visible when you open the file normally, which is exactly why it gets shared by accident.
This guide covers where to look on your own devices, how to see everything a file carries in one place, how to read the results and what a metadata viewer can and cannot tell you.
What can your computer and phone already show you?
Each platform has a viewer for the file types it cares about. They are worth knowing, and each one shows only part of the picture.
| Where | How to open it | Good for |
|---|---|---|
| Windows File Explorer | Right-click the file → Properties → Details | Photo camera, date taken and GPS; basic audio and video tags |
| macOS Preview | Open the image, then Tools → Show Inspector | EXIF camera settings and GPS |
| iPhone and Android galleries | Swipe up on a photo, or tap the info button | Date, device, lens and location on a map |
| PDF reader | File → Properties (Ctrl+D, or Cmd+D on a Mac) | Title, author, subject, keywords, creating app and dates |
| VLC | Tools → Media Information | Tags such as title, artist and date, plus codec details |
The gaps are predictable. None of these decode C2PA Content Credentials or the prompt text that Stable Diffusion tools save in PNG files. A PDF’s Properties dialog does not list the names attached to review comments, the document ID or older saved revisions still inside the file. And they each cover their own formats: a photo viewer will not tell you that a GoPro clip carries a GPS telemetry track.
How do you see everything in one place?
- Open the metadata viewer and drop in one file or several: JPEG, PNG, WebP, GIF, AVIF, HEIC, PDF, MP3, WAV, FLAC, M4A, OGG, Opus, MP4, MOV, M4V, WebM, MKV or AVI.
- Read the one-line verdict on each file: AI data found, a count of fields, or Clean when there is nothing to report.
- Open the groups. Each row shows a field, its value and a small badge naming where it was stored, such as EXIF, XMP, IPTC, PNG text, PDF info or Container.
- Use Copy report to put a plain-text summary on your clipboard for a ticket, an email or your own records.
Images and PDFs are parsed directly. Audio and video are read by a roughly 10 MB ffmpeg.wasm engine that downloads the first time and is cached after that. Nothing is uploaded: open DevTools → Network while you drop a file and you will see no request carrying it.
Inspect a file’s hidden metadataWhat does each group mean?
The viewer sorts findings by what they reveal, not by the format they came from, so a GPS tag in a JPEG and a location string in an iPhone video land in the same place.
- AI & provenance. Anything that says how a file was made by software: a decoded C2PA manifest (signer, app, actions, ingredients), IPTC digital source type and AI fields, generator tags and embedded prompts. A "Provenance reference" row means the XMP points to Content Credentials, sometimes stored online.
- Location. GPS coordinates in photos, ISO 6709 location strings in phone videos, and place names such as city and country that photo software or a newsroom typed in. This is the group to check before you post anything taken at home.
- Device. Camera or phone make and model, lens, and serial numbers. A body serial number links every photo from the same camera to each other, even when nothing else does.
- People. Author, artist, by-line, copyright and owner fields, plus the names on PDF review comments. Templates and shared computers often fill these with someone else’s name.
- Software & dates. The app and version that wrote the file, encoder tags, and creation and modification times, which can carry your time zone.
- Other. Titles, captions and keywords, embedded previews, document IDs, app-private data, chapters, cover art and data tracks such as GoPro telemetry.
Under the groups, a Kept on purpose list names things that stay even after a full clean because the file needs them to display correctly: a color profile, a JPEG orientation flag or a video’s rotation.
What should you look for in each kind of file?
| File | Check first | Why |
|---|---|---|
| Phone or camera photo | Location, Device | GPS can pinpoint a home; a serial number ties photos together |
| AI-generated image | AI & provenance | Content Credentials name the service; PNG chunks can hold your full prompt |
| People, Other | Author and comment names, and earlier revisions or a document ID that links versions | |
| Phone video | Location, Software & dates | Recording place and time, plus phone model and software version |
| Action-camera clip | Other, Location | A GPS telemetry track can log a whole route, not just a point |
| Audio file | People, Other | Names in artist and comment tags, cover art, encoder and private frames |
What can a metadata viewer not tell you?
- Whether a file is AI-generated. It shows what a file says about itself. It does not analyze pixels or sound and cannot see invisible watermarks such as SynthID, so an empty AI group proves nothing.
- Whether the metadata is true. Most fields can be edited by anyone. A C2PA manifest is signed, but the viewer reads it without validating the signature; use a C2PA validator for that.
- What is visible in the content. A street sign, a house number, a reflection or a voice saying a name are part of the file’s content, not its metadata.
- What the file name gives away. Names like
IMG_2041are harmless; names likeSmith_divorce_draft3.pdfare not, and no metadata tool reads them for you.
What if you find something you want gone?
Removal is one click away in the same tool. Everything mode strips every field while keeping what the file needs to display; AI & provenance only removes AI labels, Content Credentials and prompts and leaves the rest. The cleaned copy is inspected again before you download it, so you can see what is left, and the original on your disk is never changed.
For a specific job, the dedicated pages start in the right mode: the AI metadata remover for AI labels, the PDF metadata remover for documents, and the video and audio metadata removers for recordings. For a quick GPS strip on an everyday photo, the EXIF remover does that one job.
Frequently asked questions
Does viewing metadata change the file?
No. Reading is passive: the file is opened in your browser, parsed and listed, and nothing is written back to it. Even when you choose to remove metadata, you get a new copy to download and the original stays as it was.
Why does the viewer find nothing in my file?
Some files genuinely carry none. Screenshots hold very little, and many apps and social networks strip metadata when they save or process a file, so something downloaded from a social feed is often bare. Files straight from a camera, phone or editing app usually show plenty.
Can I check a file someone sent me?
Yes, and it is a good habit with attachments. Files sent by email, chat attachment, AirDrop or a cloud link usually arrive exactly as they left the sender’s device, metadata included, unlike photos posted publicly on most social networks.
Is there a file size limit?
Images and PDFs can be up to 512 MB. Audio and video are read without loading the whole file into memory, so large recordings can be inspected; removing their metadata usually rewrites the whole file, which a browser can do up to about 1.8 GB.