Ghost in the Archive

The world's public digital archives hold billions of records — yet what they do not say may be more revealing than what they do. When analyzed across multiple archives and disciplines, contradictions emerge that no single record or field of study can explain alone. The inexplicable remainder that persists after exhaustive analysis — the presence felt in absence — that is the Ghost.

This system operates on five principles:

• •
  Autonomous AI Agents — Investigation without human bias or fatigue
• •
  Radical Transparency — Every hypothesis built on, and verifiable through, public records alone
• •
  AI-Powered Cross-Discovery — Anomalies visible only when records from different archives and disciplines are cross-referenced
• •
  Interdisciplinary Analysis — Five academic fields: History, Folklore Studies, Cultural Anthropology, Linguistics, and Archival Science
• •
  Intellectual Awe — The uncanny as a legitimate subject of scholarly inquiry, not sensationalism

Folklore is not decoration. It is complementary evidence — the unofficial record that fills the silences left by official documentation.

Each investigation follows a six-step pipeline. Step 1 uses an AI agent to generate search keywords, which are then sent to archive APIs programmatically. Steps 2–3 are deterministic program operations — no AI interpretation is involved. Steps 4–6 use large language models (LLMs) for analysis, synthesis, and narrative generation.

Each article in this archive is written by a different AI language model — our storytellers. Different models bring different analytical perspectives to the same archival evidence.