The digitization of historical archives has expanded public access to Qajar photographs, yet visual availability does not necessarily provide historical access to the people represented. This study examines how enslaved persons and domestic dependents in Qajar photographs are discovered, named, described, and contextualized online. It employs path-based archival netnography combined with critical analysis of images, metadata, and digital affordances. A bilingual Persian-English search, domain-restricted retrieval, and examination of Women’s Worlds in Qajar Iran and selected institutional collections produced a mapping corpus of fifteen records, from which ten records were selected for in-depth analysis. The findings show that record titles generally foreground royal figures, elite women, or family names, while enslaved and dependent persons are described through terms such as slave, kaniz, servant, attendant, or “name unknown.” Explicit identification of legal status does not always recover personal identity; conversely, the preservation of a name may coexist with the concealment of enslavement under occupational terminology. Digital recirculation can reproduce incomplete metadata across platforms, while critical recaptioning can turn archival erasure itself into an object of inquiry. The study distinguishes five forms of archival visibility: contextualized visibility; legal status without individuality; named individuality without legal clarity; relational visibility; and visibility through an imposed or racialized name. It concludes that visual marginality may be extended into metadata and algorithmic marginality. A multilayered model of responsible redescription is proposed, combining the original label, critical contextualization, source of attribution, degree of certainty, searchable inscriptions, and links to album context and related scholarship.