Index and search documents with HybridRAG — a hybrid text + pixel (vision) retrieval engine. Use when the user wants to ground answers in a document corpus, build a searchable index, ingest text/HTML/PDFs/web pages, or compare text vs vision vs hybrid retrieval. Works over the `hybridrag` MCP server (tools prefixed `hybridrag_`) or the `hybridrag` CLI.
2026-07-06