Tokenize
Normalize the question into searchable terms and remove common stop words.
back to all labs
ExperimentLocal retrieval mode
RAG Retrieval Visualizer
Ask a question about the portfolio corpus. The local retrieval endpoint tokenizes the query, scores every chunk, and returns the highest-ranking context with transparent matching evidence.
01 / query
Ask the document set
Uses a fixed public corpus on this site. No embeddings or external AI provider are called.
48/30002 / retrieved-context
Ranked context chunks
Run retrieval to see which document chunks would be passed to an LLM.
Retrieval pipeline
query → tokenize → score chunks → rank → top context
Score and rank
Measure term coverage and density across every available chunk.
Assemble context
Select the highest-ranking chunks for a later generation step.
Related technical article
RAG on Postgres You Already Have →