What is the difference between Qdrant’s Retrieve Documents mode and the Answer Questions With Vector Store node?

Below is a concise explanation of the differences:

• Retrieve Documents (For Agent/Chain) Mode:
  This mode is designed to perform the core task of running a similarity search on your Qdrant collection based on the provided prompt. It returns a list of the most similar documents along with their metadata (and similarity scores). In this case, your workflow must include an additional step (or node) that takes those retrieved documents and uses them appropriately (for example, passing them into an AI model configured to answer your question).

• Answer Questions With Vector Store Node:
  This node is essentially a higher-level abstraction that not only performs the document retrieval from Qdrant but also incorporates additional logic to convert the retrieved content into a final answer. It combines the retrieval step with prompt engineering and chain processing so that the answer is produced in one node. It’s meant to simplify the workflow when you want an end-to-end answer from your vector store without manually bridging the retrieval and answer generation steps.

In summary, if you use the basic Retrieve Documents mode you have fine-grained control over how you handle the results (passing them to another agent node), whereas the “Answer Questions With Vector Store” node streamlines the process by integrating retrieval with answer generation.