SVQ.AI

· demo

An LLM-powered RAG system using LightRAG hybrid retrieval, tool-use orchestration, and FastAPI. Supports multi-format document ingestion (PDF, Word, HTML) and delivers sub-second query responses across configurable data sources.

Built for regulatory compliance teams who need instant, reliable answers from dense document corpora. The hybrid retrieval layer combines dense embeddings with graph-based entity linking, substantially reducing hallucination on domain-specific regulatory language.

Stack: FastAPI, Next.js, PostgreSQL, MinIO, Keycloak, OpenAI API, Docker, Hatchet for async pipeline orchestration.

← back to projects