RAG (retrieve)

RAG (retrieve) searches a vector knowledge base for the chunks most similar to a query and returns them. It’s the “read” half of RAG (the other is RAG (ingest), which loads the content).

It has dual use:

Connected to the rag port of an AI Agent: the agent retrieves context and injects it into the prompt before answering (so it answers with your content and doesn’t make things up).
As a standalone node in a flow: it returns the results in data.results for you to use afterwards.

Configuration

Field	Type	Required	Description
namespace	text	Yes	Namespace to search in (the same you used when indexing).
top_k	number	No	How many chunks to retrieve. Default `4`.
provider	select	Yes	Embeddings provider (openai, local, together, mistral).
model	text	Yes	Embeddings model (the same as when indexing).
credentials_id	credentials	Yes	Embeddings provider credential.
attach_to	text	No*	Agent alias (only if used as a RAG sub-node).

Example — Agent that answers about your documentation

This agent answers based on a manual previously indexed in the product-manual namespace.

RAG (retrieve):
- namespace: product-manual
- top_k: 4
- provider: openai, model: text-embedding-3-small
- credentials_id: your credential
Connect its output to the agent’s rag port (amber).
AI Agent — system_prompt: “Answer using only the retrieved context. If it isn’t there, say so.”

For “how do I configure a webhook?”, the agent retrieves the 4 most similar chunks from the manual and adds them to the prompt before the LLM writes the answer.

Output (standalone use)

{
  "query": "how do I configure a webhook",
  "results": [
    { "text": "To create a webhook…", "score": 0.89 },
    { "text": "The endpoint receives…", "score": 0.82 }
  ]
}