Running a RAG system over 11M characters of classical Buddhist texts —
   one natural defense against poisoning is that canonical texts have
  centuries of scholarly cross-referencing. Multiple independent
  editions (Chinese, Sanskrit, Pali, Tibetan) of the same sutra serve as
   built-in verification. The real challenge for us is not poisoning but
   hallucination: the LLM confidently "quoting" passages that don't
  exist in any edition.

  Building FoJin (https://fojin.app) — an open-source platform that
  aggregates Buddhist digital texts from 440+ global sources (CBETA,
  SuttaCentral, BDRC, SAT, 84000, GRETIL, etc.) into one searchable
  interface.

  Stack: React + TypeScript, FastAPI, PostgreSQL with pgvector,
  Elasticsearch 8 (ICU analysis for CJK), Redis. Features include a
  knowledge graph (9,600+ entities), 237K dictionary entries, full-text
  reading of 4,488 fascicles, and RAG-powered AI Q&A grounded in 11M
  characters of canonical text.

  GitHub: https://github.com/xr843/fojin