Metadata Remover

Privacy-focused metadata removal that runs entirely in your browser with WebAssembly. Files are processed locally and are never uploaded.

Live demo: https://ianonymous3000.github.io/metadata-remover/

Supported Formats

• JPEG - removes APP/JFIF, EXIF except minimal orientation, XMP, IPTC, Content Credentials/JUMBF, comments, and trailing appended data
• PNG - removes text, EXIF, Content Credentials/JUMBF chunks, timestamps, physical-resolution data, and unknown ancillary chunks
• WebP - removes EXIF, XMP, Content Credentials/JUMBF chunks, and unknown non-visual RIFF chunks
• AVIF - removes identifiable EXIF, XMP, and Content Credentials metadata from ISO-BMFF image containers
• GIF - removes comments, XMP blocks, and unknown application extensions
• HEIC/HEIF - removes identifiable EXIF, XMP, and Content Credentials metadata from ISO-BMFF image containers
• TIFF - removes EXIF, GPS, XMP, IPTC, camera, software, author, timestamp, physical-resolution, and copyright tags from classic TIFF IFDs
• SVG - removes metadata elements, title/description text, XML comments, processing instructions, active content, event handlers, external references, and common editor namespace attributes
• MP4/MOV - removes QuickTime/MP4 user-data, Content Credentials boxes, metadata boxes, and recording timestamps while preserving media tracks
• MP3 - removes ID3v1/ID3v2, APEv2, and Lyrics3 tags, including title/artist/album/comment frames, private frames, and embedded artwork, while preserving MPEG audio frames
• FLAC - removes Vorbis comments, embedded artwork, application blocks, cue sheets, and prepended/appended ID3 and APE tags while preserving stream info, seek tables, and audio frames
• WAV - removes LIST/INFO tags, Broadcast Wave bext, iXML/XMP, ID3 chunks, cue labels, and other non-audio chunks while preserving format and sample data
• OGG - removes Vorbis comments from Vorbis and Opus streams, including tags, vendor strings, and embedded artwork comments, while preserving codec headers and audio packets
• PDF - removes info dictionaries, XMP metadata streams, document IDs, app-specific metadata, and embedded file attachments
• DOCX - removes package properties, comments, review authors, revision IDs, and tracked-change metadata
• PPTX - removes package properties, comments, comment authors, threaded comments, and modern author parts
• XLSX - removes package properties, comments, threaded comments, person records, and revision trails
• ODT/ODS/ODP - removes ODF package metadata and annotations
• EPUB - removes OPF author, publisher, subject, description, identifiers, calibre metadata, metadata links, and reading-position sidecars while preserving book content

Cleaned files are validated as the same file type, re-scanned before download, and flagged if removable metadata remains or verification is limited by preserved embedded content. Browser-renderable images can also prepare a local visual verification preview after cleaning when you open the file details.

Use

1. Drop files into the browser.
2. Review detected metadata.
3. Click "Remove All Metadata".
4. Download cleaned files individually or as a ZIP.

All processing happens client-side in a browser worker. The app has no analytics and sends no file data to a server.

Development

Prerequisites:

• Node.js 22 or newer for the JavaScript checks and static server commands.
• The Rust toolchain in rust-toolchain.toml, including wasm32-unknown-unknown, rustfmt, and clippy.
• wasm-pack 0.14.x and Binaryen wasm-opt for release WASM builds. CI currently pins wasm-pack 0.14.0 and Binaryen 129.
• cargo-audit if you want to run the same Rust advisory check as CI.

npm run build
npm test
npm run dist
npm run serve

For a full local release check, also run:

(cd wasm && cargo fmt --check)
(cd wasm && cargo test --locked)
(cd wasm && cargo clippy --locked --all-targets -- -D warnings)
(cd wasm && cargo audit --deny warnings)
npm test
npm run dist

Rust tests include a real-encoder fixture corpus (wasm/tests/fixtures/real-device). Set METADATA_REMOVER_FIXTURE_DIR to a local folder of your own files to run them through the same clean-and-verify pipeline without committing them.

The GitHub Pages deployment cannot apply Cloudflare Pages or Netlify _headers rules. The hosted demo uses HTTPS from GitHub Pages, an in-document CSP for same-origin app resources, and a no-referrer meta policy. Header-only protections such as HSTS, frame-ancestors, COOP/COEP/CORP, X-Content-Type-Options, and Permissions-Policy require a host or proxy that can set HTTP response headers.

Limitations

• The browser UI rejects individual files over 100 MB and applies a device-memory-aware batch budget. Download or clear finished files before adding more.
• JPEG orientation and color-transform/profile segments are preserved to avoid sideways images and color shifts.
• BigTIFF is not supported; save or export as classic TIFF first.
• AVIF, HEIC/HEIF, and MP4/MOV cleaning target known ISO-BMFF metadata structures this parser can identify.
• EPUB cleaning preserves required package fields by normalizing identifiers and modified timestamps to neutral values.
• OGG cleaning supports single-stream Vorbis and Opus files; multiplexed (e.g. video) and chained OGG streams are rejected. Encoder signatures embedded inside audio frames (such as LAME headers) are preserved.
• Embedded images inside Office, ODF, and EPUB packages are recursively cleaned when their content is a supported image format; unreadable or unsupported embedded files are preserved, flagged, and keep the result in review.
• Visual verification is limited to browser-renderable JPEG, PNG, WebP, AVIF, GIF, and cleaned SVG previews, and runs after cleaning when local memory headroom allows.
• SVG active content and external references are removed, which can change interactive or externally styled SVGs.
• SVG base64 data URIs holding supported raster images are decoded, recursively cleaned, and re-encoded; other data URI content (CSS references, non-base64 payloads, nested SVG) is preserved and keeps the result in review.
• Unknown/private TIFF tags and uncommon or ftyp-less video variants may be preserved or rejected.
• Legacy binary Office files (.doc, .xls, .ppt) are not supported; save them as .docx, .xlsx, or .pptx first.
• This tool removes structural metadata; it does not detect steganographic content embedded in pixels or document bodies.

License

MIT