# Metadata Check > Drop photos, PDFs, Office files, MP4 and HAR at once; see what each carries: location, author names, hidden content, tokens. Clear in batch. Nothing uploaded. - URL: https://bytescope.dev/en/metadata-check/ - Locale: en · Also: [ja](https://bytescope.dev/ja/metadata-check.md) · [zh-Hant](https://bytescope.dev/zh-Hant/metadata-check.md) - Runs: Your files never leave your browser — all processing is local. - Requires: Any modern browser (Chrome, Edge, Firefox, Safari), desktop and mobile. Writing the cleaned files straight into a folder needs Chrome, Edge or Opera on desktop (File System Access API); everywhere else the output is a ZIP download. - Updated: 2026-09-08 ## What it does Before a bundle of files goes to a lawyer, a client, a journalist or a marketplace, the question is rarely “which tool handles a PDF” — it is “what is in this pile”. This page takes the pile. Drop files or a whole folder; each one is identified by its bytes rather than its extension and handed to the same inspector the format's own page uses: JPEG, PNG, WebP, HEIC and AVIF go to `/remove-exif`, PDF to `/pdf-metadata`, Word, Excel and PowerPoint to `/office-inspector`, MP4 and MOV to `/remove-video-location`, and a `.har` export from DevTools to `/har-sanitizer`. The findings are normalised into one list per file — location, people, credentials, device, time, software, hidden content, paths, Content Credentials — with a summary at the top, such as “12 files: 4 with location, 6 with an author name, 1 with tracked changes, 2 this tool cannot check”. You can filter to only the files with location, or only the ones that could not be checked. Nothing is uploaded; the work runs in a worker inside your browser. ### What each format is checked for Photos: GPS, the artist and copyright fields, camera make, model and serial number, capture time, software, the embedded thumbnail, XMP and IPTC blocks, and any data appended after the end of the image. PDF: author, creator and producer, creation and modification dates, title, subject, keywords and XMP. Word, Excel and PowerPoint: core and app properties, the people named in comments and tracked changes, template and local paths, the thumbnail, custom XML, hidden text, hidden sheets and rows, speaker notes, cropped-away image areas, macros and more. Video: location only — the `©xyz`, `loci` and QuickTime keys where iPhones and cameras write coordinates, plus a Content Credentials marker; a video's title, camera model or capture time is not read here. HAR: cookies, `Authorization` headers, tokens, JWTs, private keys, possible personal data and internal hosts. Each card lists which of these categories were actually checked for that file, so an empty card reads “nothing found among the checked types”, never “nothing in it”. ### Clearing in batch The clear step calls the same function with the same defaults as each format's own page — not a second, looser set of rules. Photos are stripped without re-encoding, so the pixels stay byte-identical. PDFs have the Info dictionary and XMP cleared. Office files lose properties, thumbnail, custom properties and absolute paths, while tracked changes, comments and the names inside them are reported but left alone, because rewriting those alters the document — `/office-inspector` has the options for that. Videos have their location atoms blanked in place. HAR files get the specialist page's default selection. Content Credentials (C2PA) are one checkbox for the whole batch and are kept by default: clearing them removes the provenance signature that says who made the file. Output is a single ZIP or, in browsers that support it, written straight into a folder you pick; the originals are never modified. A file with nothing to clear is listed as such and left out of the output. The ZIP takes files up to 256 MiB each and 512 MiB in total; anything larger is offered as a separate download instead. ### What this page does not claim The summary says what was found. It never says the files are safe, clean or ready to send, and neither does the result after clearing. Two phrases mean different things here: “cannot check” is a format this page has no inspector for — an MP3, a ZIP, a PSD — and “cannot determine” is a supported format that could not be parsed: truncated, encrypted, or structured in a way the parser does not recognise. Neither is shown as a blank result or a green tick. Some things are not checked at all, on any page: what is visible in the picture or in the video frames themselves (a street sign, a screen, a face), the file name, files inside archives (a ZIP or RAR is identified, not opened), email files (`.eml`, `.msg`), audio tags (ID3 in MP3, Vorbis comments in FLAC) and encrypted files. After clearing, every output is checked again with the same inspector and shown next to the original, so you can see what changed — the final read before sending is still yours. ## Limits Does not write a parser of its own — each format goes to the same inspector as its own page, and a format with no inspector is reported as one this tool cannot check. It does not look at what is visible in a picture or a video frame, at file names, inside ZIP or RAR archives, at .eml or .msg email files, at MP3 or FLAC audio tags, or inside encrypted files. In video it looks for location only. It never says a file or a batch is safe or clean — it lists what it found and which types it checked. ## FAQ ### How do I check a file for metadata when I don't know what format it is? Drop it here. The page reads the first and last bytes of each file and identifies the format from its content, not its extension, so a `.jpg` that is really a PNG or a renamed PDF is still routed to the right inspector. If the format is one the page cannot check, it says so and lists the formats it can: JPEG, PNG, WebP, HEIC, AVIF, PDF, Word, Excel, PowerPoint, MP4, MOV and HAR. ### What metadata is in my file? It depends on the format, and the card for each file tells you exactly which types were checked. A photo can carry GPS coordinates, the camera's make, model and serial number, the capture time, an artist name and an embedded thumbnail. A PDF carries author, creator, producer and dates. A Word, Excel or PowerPoint file carries the people named in comments and tracked changes, template paths, hidden text, hidden sheets and speaker notes. A video from a phone usually carries only the coordinates — and that is all this page looks for in video. A HAR file carries the cookies and tokens of your logged-in session. ### Are my files uploaded anywhere? No. Identification, inspection and clearing all run in a web worker in your browser; the ZIP you download is built from memory. There is no API behind this page, which is the point for exactly the files you are checking because you are about to share them. ### Can I remove metadata from many files at once before sending them? Yes. After the check, one button clears every supported file using the same function and defaults as that format's own page, then packs the results into a ZIP or writes them into a folder you choose. The originals are not touched. Every output is checked again and shown next to the original, so you see what was removed and what remains. Content Credentials (C2PA) are kept unless you tick the batch-level box to clear them. ### If nothing is found in a folder, does that mean the files contain no personal information? No. It means nothing was found among the types this page checks, and each card lists those types. The page does not look at what is visible in a photo or a video frame, at file names, inside ZIP or RAR archives, at `.eml` or `.msg` email files, at MP3 or FLAC audio tags, or inside encrypted files. A format it cannot check is reported as such, never as a clean result. Read the files yourself before they go out. ### What is the difference between “cannot check” and “cannot determine”? “Cannot check” means the format has no inspector here — an MP3, a ZIP, a PSD, or a file whose format could not be identified with confidence. “Cannot determine” means the format is supported but the file could not be parsed: a truncated JPEG, an encrypted PDF, a damaged Office file, a video with no readable `moov` box. Both stay visible in the summary, so a count of zero findings never hides a file that was never actually inspected. ### Does it check videos for everything a photo is checked for? No. For MP4 and MOV the page only looks for location — the `©xyz`, `loci` and QuickTime metadata keys where phones and cameras write coordinates, including the small `.MOV` that comes with an iPhone Live Photo — plus a Content Credentials marker. Creation time, camera model or title inside a video are not read or cleared here, and the card says so. ### Why are Content Credentials kept by default? Content Credentials (C2PA) are a signed record of who made the file and how it was edited; the same signature is what lets someone verify a photo is genuine. Clearing them removes that proof, which is the right call for some sends and the wrong one for others, so it is a single checkbox for the whole batch, off by default. The same rule applies on `/remove-exif` and `/remove-video-location`. ## Related tools - [Remove EXIF](https://bytescope.dev/en/remove-exif.md) — Strip EXIF, GPS and XMP from photos, with more options per file. - [Remove PDF Metadata](https://bytescope.dev/en/pdf-metadata.md) — See and clear a PDF's author, dates and XMP field by field. - [Office Document Inspector](https://bytescope.dev/en/office-inspector.md) — Tracked changes, comments, hidden sheets and paths in Office files. - [Remove Video Location](https://bytescope.dev/en/remove-video-location.md) — Find and blank the GPS location in iPhone and camera videos. - [HAR File Sanitizer](https://bytescope.dev/en/har-sanitizer.md) — Replace the cookies and tokens in a HAR export before you share it.