---

name: knowledge-rag description: "Owns KnowledgeBase workflows: loaders, splitters, embeddings, vector DBs, OCR, and retrieval tooling." disable-model-invocation: true metadata: disco-role: operating license: MIT

knowledge-rag

Use this route for document ingestion, chunking, embedding, vector search, OCR, and KnowledgeBase workflows.

Include

• KnowledgeBase construction and retrieval/search helpers.
• Loader, splitter, embedding, OCR, and vector DB selection.
• Document lifecycle methods such as add, refresh, update metadata, and retrieval tooling.

Exclude

• Chat/session persistence → chat-memory-storage
• Core model/provider selection → models-and-providers
• Core agent execution → agent-runtime

Start here

• references/knowledge-base-workflows.md
• references/backends-and-extras.md
• references/troubleshooting.md
• scripts/check_rag_optional_imports.py