Last year, a user rediscovered 1,351 physical photos at their grandmother’s house, spanning decades of family history. After organizing them by film stock and aspect ratio, they collaborated with their grandmother to contextualize a wedding album, uncovering forgotten stories. This sparked an experiment using MediaWiki to create a digital archive, linking photos to verified events and people.

The project evolved beyond static images. For a 2012 trip to Coorg, the user input 625 photos into Claude Code, which generated a detailed wiki page by analyzing timestamps and visual cues. Without GPS data, the model identified locations, transportation modes, and even recognized people after manual input. Later, a Mexico City trip yielded 291 photos and 343 videos with EXIF metadata. By cross-referencing bank statements, Uber trips, and Shazam playlists, Claude Code mapped dining experiences, soccer matches, and background music, creating a narrative-rich archive.

The tool’s capabilities expanded when the user integrated social media data—100k messages and voice notes from WhatsApp, Facebook, and Instagram. Claude Code traced friendship timelines, life events, and shared memories, producing pages that felt personally written. This shift from family history to social archiving revealed the power of combining AI with encyclopedic frameworks.

The experiment underscores how tools like MediaWiki and language models can preserve ephemeral memories. By merging oral history, metadata, and crowdsourced data, individuals can transform fragmented digital footprints into cohesive, searchable narratives—bridging gaps in collective memory.