Images & photos · Image Metadata Privacy Tool
From IPTC IIM to Content Credentials: A Timeline of Photo Metadata
· Background
exif image-privacy file-formats
Photo metadata grew in layers: newsroom captions in 1991, camera settings in 1995, PNG text chunks in 1996, Adobe's XMP in 2001, industry reconciliation in the late 2000s, and signed provenance manifests in the 2020s. Each layer is still found in files today.
Why one photo can carry five kinds of metadata — the concrete result of thirty years of standards layered on top of each other
One photograph can carry camera directories, newsroom resources, free-text chunks, editing packets and newer provenance structures at the same time. The workbook presents a historical timeline, but the repository does not contain standards sources supporting its dates. This article therefore uses the sequence as a map of layers and states exactly which ones the shipped cleaner handles.
IPTC IIM as the APP13 layer this cleaner removes
APP13 Photoshop/IPTC is a JPEG metadata layer the classifier removes as a complete segment. The implementation does not parse IIM fields or establish the workbook’s date, organisation history or newsroom adoption claims. If those records are operationally important, preserve the master and manage them in an IPTC-aware system before creating a stripped derivative.
EXIF as the TIFF-directory layer this reader parses
EXIF is the layer this reader understands most deeply: a TIFF header leads to IFD0 and then bounded Exif and GPS sub-directories. The same parser serves JPEG APP1 Exif, PNG eXIf and WebP EXIF. Whole-block removal also takes unread MakerNote data and any embedded EXIF thumbnail without requiring every internal value to be decoded.
PNG text and eXIf chunks in the supported container
PNG can carry tEXt, zTXt, iTXt, tIME and eXIf chunks. The tool displays plain tEXt, names compressed or international text by keyword and parses eXIf through the TIFF reader. It removes those chunks while retaining colour, density, transparency and compressed IDAT data. The repository does not verify the workbook’s format-history dates.
XMP as a detected but unparsed packet
XMP is detected separately from EXIF and removed from JPEG and WebP, but its XML is not parsed into history or caption fields. That distinction prevents the presence of a packet from being mistaken for knowledge of its contents. PNG XMP is not named as a separate supported chunk, so the article does not generalise XMP removal beyond the matrix.
Metadata reconciliation is outside this tool
The workbook names an industry reconciliation period, but metadata synchronisation and working-group history are not represented in this codebase. ToolAcre does not reconcile duplicate captions, choose authoritative values or migrate fields between schemas. Its operation is intentionally simpler: report readable data and drop recognised container blocks from a copy.
Content credentials are outside this tool
Content Credentials and C2PA require specialised manifest and signature handling not found in the support matrix, parser or stripper. The tool makes no claim to inspect, preserve or remove those systems, and this article omits the workbook’s date assertion. Use credential-aware software when provenance verification or signed assertions are part of the delivery.
Takeaway and limits: what a stripper addresses — the Image Metadata Privacy Tool targets EXIF, GPS, IPTC and XMP; check its page for how newer provenance blocks are handled
The useful timeline ends at an implementation boundary: ToolAcre parses selected EXIF directories, detects or removes specified IPTC, text and XMP blocks, keeps rendering data and refuses unsupported containers. Review what the result says was removed and kept, preserve provenance-bearing originals and never translate “no readable fields” into “every metadata layer is gone.”