S

Structured Document Retriever

Project information

A modular Retrieval-Augmented Generation (RAG) system for textbooks and long-form documents. Uses layout-aware heuristics + LLM-verified chunking to generalize across document formats (no hardcoded regex), stores embeddings and content in PostgreSQL/pgvector, and answers questions with query-routing guardrails (greeting/in-domain/off-topic/blocked detection).

RAG pgvector PostgreSQL embeddings document-chu... Python

33 Commits
1 Branch
0 Tags

README

Created on

July 11, 2026

Loading