RAG — search that grounds the answer

A language model runs in your browser and searches by meaning for the right fragments of a migrant storytelling archive to ground the answer.