Tiny Shakespeare RAG

Retrieval runs entirely offline with @ktjn/searchable against a static index of 1,422 Shakespeare dialogue passages. Generation runs entirely in this tab with @mlc-ai/web-llm over WebGPU — no server, nothing sent anywhere.

Model

Chat

Source text: tiny-shakespeare (public domain, concatenated Shakespeare excerpts, no act/scene attribution in the source file). Index built with the Python searchable indexer; retrieval runs with the TypeScript @ktjn/searchable client. See examples/rag-webllm/README.md for how to rebuild it.