Titanic Historical RAG

The killer feature: contradictions, not consensus

Most RAG systems try to return one answer. The witnesses in the Titanic inquiries gave conflicting accounts on almost everything — speed, lifeboat counts, ice warnings, who gave which order. This tool embraces that. Ask “How many people were in Ismay’s lifeboat?” and you’ll see Ismay’s “about 45” set next to Officer Lowe’s “only twelve”, with a confidence-scored explanation of why the statements conflict.

Architecture

What gets indexed

• US Senate Inquiry (1912) — 1,173 pages, fully ingested. ~40,000 chunks across 68 witnesses.
• British Board of Trade Inquiry — format sample only; full ~2,200-page transcript on the roadmap.

Long-term goal: ingest the full British Inquiry to surface cross-inquiry contradictions — where the same witness testified differently in the two proceedings.

API shape

POST /search/contradictions returns a list of pairwise contradictions with witness names, the two conflicting claims, a 0–1 confidence score, a one-sentence explanation, and the source chunks behind each claim.

{
  "witness_a": "Joseph Bruce Ismay",
  "witness_b": "Harold Godfrey Lowe",
  "claim_a": "Ismay's lifeboat had approximately forty-five people",
  "claim_b": "Ismay's lifeboat (Boat C) had only twelve people",
  "confidence": 0.95,
  "explanation": "The two witnesses provide directly conflicting specific counts: forty-five versus twelve, which cannot both be true."
}

Roadmap

• Done — Killer feature shipped end-to-end and deployed at titanic.higuera.io (search → LLM verdicts → side-by-side UI behind a custom domain).
• Next — Hook /witnesses to the canonical witness index; retire legacy ChromaDB code.
• Soon — British Inquiry pipeline (speaker-tag parser, separate witness index, ingest full corpus).
• Production — DynamoDB-backed verdict cache, rate limiter on the contradictions endpoint.