# H/M Photo Privacy & Provenance Scanner

Canonical URL: https://hogarmas.net/content-credentials/metadata/

## Purpose

Inspect readable photo metadata before sharing a file. The workbench groups EXIF, GPS, device, date, creator, software, caption, embedded-preview and technical fields. It does not decide whether media is true, original or AI-generated.

## Execution boundary

- Source files are processed in the browser.
- Source files, filenames, metadata values, GPS coordinates and device identifiers are not uploaded.
- Browser storage is not used.
- SHA-256 is computed locally with Web Crypto.
- Fixed aggregate action IDs and dwell time are the only application metrics.
- Automated browser sessions identified by `navigator.webdriver` are excluded from metrics.

## Supported input

One image up to 50 MB. The interface accepts JPEG, PNG, WebP, HEIC, HEIF, TIFF and AVIF. Browser preview and clean-copy support depend on the browser codec. The pinned exifr 7.1.3 parser reads supported EXIF, TIFF, GPS, XMP, IPTC, ICC, JFIF and PNG IHDR fields.

## Workflow

1. Select or drop one local image.
2. Review category counts while sensitive values remain hidden.
3. Explicitly reveal values only when needed.
4. Filter by tag name or category.
5. For a standard-orientation JPEG, optionally remove APP and COM metadata segments before the first image scan without recompressing its encoded image scan. Other supported previewable formats create a local Canvas-rendered JPEG, PNG or WebP share copy.
6. Re-scan the new copy and compare observed sensitive-field counts.
7. Export a privacy-minimized JSON or Markdown report.
8. Open the separate C2PA verifier for signed provenance evidence.

## Clean-copy boundary

A standard-orientation JPEG can have APP and COM segments removed before its first image scan without recompressing the encoded image scan. This removes EXIF, XMP, IPTC, ICC, comments and C2PA/JUMBF data found in those segments, but may also remove a color profile, declared orientation or Content Credentials. Other formats are locally re-encoded and may change pixels, compression, color, orientation or transparency. Re-scan results cover only metadata readable by the pinned parser and do not guarantee every proprietary container or later platform transformation.

## Report contract

- Schema: https://hogarmas.net/content-credentials/metadata/report-schema.json
- Raw metadata values are excluded.
- Filename, coordinates and device identifiers are excluded.
- Local source and clean-copy SHA-256 values may be included.
- The report does not validate C2PA signatures or authenticate an issuer.

## Related

- C2PA verifier: https://hogarmas.net/content-credentials/
- English search guide: https://hogarmas.net/image-metadata-viewer/
- Spanish search guide: https://hogarmas.net/ver-metadatos-foto/
- Chinese search guide: https://hogarmas.net/zh/image-metadata-viewer/
- Parser source: https://github.com/MikeKovarik/exifr