find documents by meaning, without hallucinating
Semantic search across one or more knowledge bases — it finds the right document by meaning, not a chatbot.
The RAGflow Search block is deliberately retrieval-only: it fetches the best-matching documents and shows them with a title, a snippet and a match score — it does not compose an AI answer, so it can’t make anything up. It’s for when you want the source, not a summary.
It searches the content of the files — scans, images and tables included (OCR, layout and image understanding) — not just titles and metadata like Moodle’s built-in search. Results can be tuned for quality (a relevance floor — fewer but better hits) and optionally sharpened with a rerank model; it searches one or several knowledge bases depending on how it’s configured.
Free and open-source (GPLv3). Requirements: Moodle 5.0–5.2 and the RAGflow AI provider (which supplies the RAGflow connection).
Support: Professional, English- and German-speaking support is available from RAGcon GmbH (www.ragcon.ai).
