What is query transformation in RAG and how does HyDE implement it?

Query transformation, or query rewriting, is a RAG optimization where an LLM rephrases the user's original query so it more closely matches the wording used in documents for vector similarity search. HyDE, short for Hypothetical Document Embeddings, implements this by generating a hypothetical answer to the query and then running the vector DB search using that synthetic text instead of the raw question.

Query transformation is needed because a vector database looks for semantic similarity, and a user's phrasing may not align with how the documents are written. To solve this, another LLM call rewrites the query into a version more likely to match useful chunks. HyDE is a popular instance of this approach: the LLM is prompted to generate a paragraph of text answering the query, and that hypothetical answer is embedded and used as the search input to the vector DB. This synthetic text can bridge wording gaps between the user and the documents. The technique can also be adapted to handle language mismatches, such as when documents are in Spanish but the question is in English, by using a prompt and LLM to translate the query first.