Enables AI agents and MCP hosts to query a local hybrid vector and BM25 index over scattered notes, docs, and chat logs, returning section-level citations for grounded answers.
Turns a user's own Markdown wiki into a locally indexed knowledge base, enabling semantic search, per-page retrieval, source listing, and citation-backed RAG question answering. All embedding and generation run through a local Ollama instance, so queries and documents never leave the machine.
Enables semantic search over local AI-agent session logs (e.g., Claude Code, Hermes) using local embeddings, returning relevant chunks without exposing full transcripts.