Hidden in every GPS coordinate
you share.

GPS, EXIF, AI tags, author names, gone in seconds. 100% in your browser. Try now below

metastrip
↑
drag & drop files or click to browse
jpeg, png, webp, pdf, docx, xlsx, pptx, mp4, mov, mp3, m4a, flac, wav
metastrip v2.0 — client-side metadata removal
batch limit: 20 files | all processing happens in your browser
metastrip❯
…/uploads❯
⎇ main↑3✱
❯
_
v2.0client-sideno tracking⎇ main↑3✱✓ build passing—
How it works

Privacy is in the details.
Specifically, the metadata.

What's hidden

Every file you share is talking.

Photos carry GPS coordinates, device serial numbers, camera info, and timestamps. PDFs carry author names, edit history, and software fingerprints. AI-generated images carry C2PA content credentials identifying which tool made them.

  • GPS to within 5 metres of where you stood
  • Device + serial number that fingerprints every photo to one phone
  • AI generation tags from DALL·E, Midjourney, ChatGPT, Firefly
  • Author names + tracked changes in Word, Excel, and PDF documents
How it works

Drop a file. See what's exposed. Strip it.

MetaStrip runs entirely in your browser. There is no upload, no server, no temporary cache. You drop the file in, the tool reads its metadata locally with libraries like piexifjs and pdf-lib, shows you exactly what's embedded, and writes a clean copy back out.

  • Files never leave your device — verifiable in DevTools
  • Batch up to 20 files at once
  • Open source under MIT. The code is on GitHub.
  • Works on mobile and desktop, online or offline
When to use it

Before you share anything you didn't write yourself.

Anywhere a file leaves your device is somewhere your metadata follows. Contracts to opposing counsel, photos to forums or marketplaces, AI-generated work to clients, journalism submissions, anything posted under a pseudonym.

  • Before posting photos taken at home
  • Before sending a Word doc with revision history to a client
  • Before submitting AI-assisted work where disclosure changes acceptance
  • Before publishing anything anonymously — devices fingerprint themselves
FAQ

Questions? Answers.

Everything people usually ask about how the tool works, what it does to your files, and what happens to your data.

Yes. Every file is processed entirely in your browser using JavaScript. Nothing is uploaded, nothing is cached on a server, nothing is sent anywhere. You can verify this yourself: open DevTools → Network tab → process a file → watch zero requests fire. The whole tool ships with the page on first load and runs locally from then on.
Photos: GPS coordinates, EXIF (camera make, model, serial number, settings), IPTC, XMP, C2PA content credentials, AI generation tags, embedded thumbnails, and timestamps. PDFs: author, creator app, producer, title, subject, keywords, custom properties, and timestamps. Word/Excel/PowerPoint: author, last-modified-by, company, tracked changes, comments, revision history, and template metadata. Video (MP4/MOV/M4V): GPS coordinates, device make/model/software, handler vendor IDs, creation and track timestamps, and tool fingerprints in udta/meta atoms. Audio (MP3/M4A/FLAC/WAV): ID3v1 and ID3v2 tags, Vorbis comments, embedded album art, RIFF LIST/INFO chunks, Broadcast Wave (bext) extensions, and iXML metadata. The full per-file-type breakdown is on the privacy page.
Images: JPEG, PNG, WebP. Documents: PDF, Word (.docx), Excel (.xlsx), PowerPoint (.pptx). Video: MP4, MOV, M4V. Audio: MP3, M4A, FLAC, WAV. HEIC, TIFF, GIF, and additional document formats are on the roadmap.
Yes. No ads, no tracking, no signup, no upload limits, no premium tier. The source is open on GitHub under MIT license. If the tool helps you, you can buy me a coffee on Ko-fi — completely optional, never required to use anything.
No. Image content, document content, and quality stay exactly the same. Only the hidden metadata is removed. File size will drop slightly because that metadata is no longer present (typically a few KB per photo, sometimes more for AI-generated images that carry large C2PA manifests).
iPhone and Android photo apps remove GPS coordinates but keep most other metadata: device model, serial number, software version, edit history, camera settings, AI generation tags, and timestamps. MetaStrip removes all of it — and works on PDFs and Office documents too, not just photos. It also shows you exactly what was embedded before you strip it, so you can see what your files were leaking.
AI tools like DALL·E, Midjourney, ChatGPT image generation, and Adobe Firefly embed C2PA 'content credentials' that cryptographically identify images as AI-generated. MetaStrip strips these alongside the rest of the metadata. Worth knowing: this removes metadata-based AI markers but does not remove pixel-level steganographic watermarks like Google's SynthID — those are embedded in the image data itself, not in the metadata, and require different techniques.
Yes — MIT licensed, source code on GitHub. You can audit the code, run it locally, fork it, or contribute. The whole tool is built with public libraries (piexifjs, pdf-lib, jszip) so you can verify exactly what it does to your files.
Yes — drop up to 20 files at once. Each is processed independently in the browser, and you can download them individually or as a single zip. Batch limits exist purely because browser memory is finite, not because we're holding back features.
Yes. The full tool runs on iOS and Android browsers. The tool stays the same, the layout adapts.

Still curious? hello@metastrip.app

Portrait of Lars Holmstrom
My name is Lars, I'm from Melbourne, Australia. I'm a cyber security graduate from the University of Tasmania, now an indie dev who likes privacy — the tool you're using right now is the tool I built for myself first.
Find me on X