I’m testing an N8N and Qdrant integration and encounter a ‘400 Bad Request’ error about a missing vector identifier. How can I properly set up the collection?
{
"workflow": "DemoProcess",
"nodes": [
{"id": "node1", "type": "Embedder", "params": {"model": "small-text"}},
{"id": "node2", "type": "RepoFetcher", "params": {"source": "demo-repo"}},
{"id": "node3", "type": "TextDivider"},
{"id": "node4", "type": "Trigger"},
{"id": "node5", "type": "VectorStore", "params": {"collection": "docs_new"}}
]
}
hey, check your vector field mapping, i had a simlar issue. you might need to add explicit id params in the collection setup in n8n before inserting docs. hope this helps.
It appears that the error stems from missing an explicit identifier in the document insertion process. In my experience, ensuring the presence of a unique ID field when configuring the vector store has been key. The solution involved setting up the mapping so that every document passed to Qdrant has this required identifier. It is worth revisiting the configuration on the VectorStore node to verify that the id field is being correctly initialized before the integration forwards data to Qdrant.
I ran into a similar issue while integrating N8N with Qdrant a few weeks ago. My initial assumption was that the node would automatically generate vector identifiers, but that wasn’t the case. I eventually discovered that the vector store node needs a dedicated field explicitly mapped to the document’s id. Indeed, after revising the collection configuration and ensuring that each document is paired with a unique identifier, the error disappeared. It might also help to check if your workflow is passing any custom parameters or if updates in Qdrant’s API have changed expected behavior. Careful review of both platform documentations proved invaluable.
hey i’ve seen this before. try manually setting the unique id for your vectors as the default might not work. its not a qdrant error but n8n needing explicit id mapping. check your node config and you should be set.