SBU RAG ingestionQdrant sharedQdrant analysis

Insert company knowledge into the board member memory.

Upload SBU documentation, import a web page, or paste source text. The backend chunks it and stores embeddings in the same retrieval pipeline used by chat.

Pipeline target

PyPDFLoader, chunking, embeddings, Qdrant

Dataset

sbu_document_upload

Kind

sbu_document_chunk

TitleSource name

Drop or select company PDF documentationPDF files are parsed on the backend before embedding.

Chat retrieval can use the new chunks immediately after Qdrant upsert completes.