A developer created django-graph-search, a library that adds a vector search layer on top of existing Django ORM data without requiring schema changes or new tables. It traverses ORM relations to build rich text documents representing objects and their context, which are embedded into vectors for semantic search. This enables AI agents to answer natural language queries over live app data using a standard Retrieval-Augmented Generation (RAG) pattern with minimal code changes. The solution supports automatic incremental indexing, configurable field weighting, multiple vector backends (ChromaDB, FAISS, pgvector, Qdrant), and embedding models (sentence-transformers, OpenAI, Cohere). It also optionally supports query expansion, reranking, and conversational memory via an optional LangGraph pipeline.
Use Case
Opening the operator briefing
Pulling the full operator breakdown, tooling context, and verification notes.
