RCLI includes local RAG out of the box. You can ingest PDFs, DOCX, and plain text, then query by voice or text: rcli rag ingest ~/Documents/notes rcli ask --ra

RCLI includes local RAG out of the box. You can ingest PDFs, DOCX, and plain text, then query by voice or text: rcli rag ingest ~/Documents/notes rcli ask --rag ~/Library/RCLI/index "summarize the project plan" It uses hybrid retrieval (vector + BM25 with Reciprocal Rank Fusion) and runs at ~4ms over 5K+ chunks. Embeddings are computed locally with Snowflake Arctic, so nothing leaves you're machine.