Class QdrantVectorStore

Qdrant vector store adapter for production vector search.

Qdrant is a high-performance vector database with filtering, payload storage, and distributed deployment support.

Requires @qdrant/js-client-rest package to be installed:

pnpm add @qdrant/js-client-rest

Example

// Local Qdrant instance
const memoryService = new MemoryService({
  storage: new VectorStorageProvider({
    vectorStore: new QdrantVectorStore({
      url: "http://localhost:6333",
      collectionName: "agent-memories",
      dimensions: 1536, // OpenAI ada-002 dimensions
    }),
  }),
  embeddingProvider: new OpenAIEmbeddingProvider(),
});

// Qdrant Cloud
const cloudStore = new QdrantVectorStore({
  url: "https://your-cluster.qdrant.io",
  apiKey: process.env.QDRANT_API_KEY,
  collectionName: "memories",
  dimensions: 1536,
});

Implements

Constructors

constructor

Methods

upsert search delete count deleteCollection getCollectionInfo

Constructors

constructor

Methods

upsert

search

  • search(params): Promise<{
        id: string;
        score: number;
        metadata: Record<string, unknown>;
    }[]>

  • Search for similar vectors.

    Parameters

    • params: {
          vector: number[];
          topK: number;
          filter?: Record<string, unknown>;
      }
      • vector: number[]
      • topK: number
      • Optionalfilter?: Record<string, unknown>

    Returns Promise<{
        id: string;
        score: number;
        metadata: Record<string, unknown>;
    }[]>

delete

count

deleteCollection

getCollectionInfo

  • getCollectionInfo(): Promise<{
        vectorsCount: number;
        pointsCount: number;
        status: string;
    }>

  • Get collection info.

    Returns Promise<{
        vectorsCount: number;
        pointsCount: number;
        status: string;
    }>

Settings

Member Visibility

On This Page

Constructors
Methods