How can get bm25 vector in local qdrant

I want to do it like in this link: Index legal documents for hybrid search with Qdrant, OpenAI & BM25 | n8n workflow template
I’m using a local qdrant and I don’t understand/can’t get bm25 vectors.

I use node qdrant query points
Query:
{
“fusion”: “rrf”
}
Prefetch:
[
{
“query”: [ {{ $json.batch[0].embedding }} ],
“using”: “qwen3_0_6b”,
“limit”: 20
},
{
“query”: {
“text”: {{ $json.batch[0].question }},
“model”: “qdrant/bm25”,
},
“using”: “bm25”,
“limit”: 20
}
]

It works without
{
“query”: {
“text”: {{ $json.batch[0].question }},
“model”: “qdrant/bm25”,
},
“using”: “bm25”,
“limit”: 20
}

Hi @g916673 Welcome!
I think qdrant documentation should answer this more clearly

Thank you very much, it was just an extra comma (“model”: “qdrant/bm25”,).

But I followed all the steps you recommended in the documentation…

1 Like

@g916673 Glad it put some help, consider marking that as a solution so future discoverers can refer from this too.

Cheers!

1 Like