RAG Search Node
Search documents via vector similarity using natural language queries within scenarios.
Node: RAG Search
To use stored data inside a scenario, connect the RAG Search node from the AI Agent → Actions category.
Main Fields (RAG Search Node)
Field | Description |
Storage | Select the storage to search in |
Question | Natural language query |
Top_k | Number of chunks to return (default: 5, max: 20) |
How It Works
- You upload a document into a storage
- The document is automatically split into chunks and indexed
- RAG Search receives a query and performs embedding-based retrieval
- The node returns raw chunks that match the query
Node Execution Example
A natural language query is passed into the node, which returns a list of matching chunks based on the specified top_k.
Did this answer your question?
😞
😐
🤩