Query Rewriter

A transformation engine that converts ambiguous user input into optimized search terms to maximize retrieval precision.

Query Rewriter functions as a strategic translation layer for RAG (Retrieval-Augmented Generation) stacks. It uses LLMs to refine raw prompts: stripping conversational filler, correcting syntax, and injecting domain-specific keywords. By routing these optimized queries to vector stores (like Pinecone or Weaviate), it solves the 'vocabulary mismatch' problem. This process typically boosts document recall by 30-50% in enterprise search applications while reducing the risk of model hallucinations.

https://python.langchain.com/docs/concepts/query_analysis/