Skip to main content

Qdrant: Embeddings

Deprecated

This function is deprecated and can no longer be added to new workflows. Existing workflows using it keep running. Use Qdrant: Upsert instead.

Store embeddings vectors into a Qdrant database collection.

If the collection does not exist yet, the node creates it automatically, using the dimension of the first input vector and Cosine distance.

Pre-requisite: Install Qdrant application Profile > {Organization} > Applications to grant Zparse access.

Parameters​

Applicationrequired

Select configured Qdrant application.

Collectionrequired
Which collection to use.
Collection mode
  • Single Embedding (default) — stores one dense vector per point. - Hybrid Search — stores both a dense and a sparse vector per point for keyword-aware search.
Dense vector name

(Hybrid Search mode only) Name of the dense vector in the collection.

Sparse vector name

(Hybrid Search mode only) Name of the sparse vector in the collection.

Sparse method

(Hybrid Search mode only) Method used to generate the sparse vector: BM25 (default) or Splade.

Input​

JSON Embeddingsrequired

Embeddings request with model like:

{
   "embeddings": [
     {
       "identifier": <String>, # Unique identifier for given claim
       "claim": <String>, # Actual text from which embeddings vector was generated
       "vector": [f32, f32, ...], # Actual vector after embeddings generation
       "payload": <Json>, # Additional metadata to join to your claim
     }
   ]
}

Output​

Triggerrequired

Trigger connector on vector successfully stored