Get started

pnpm add @anvia/core @anvia/qdrant @qdrant/js-client-rest

import { retrieveDocuments } from '@anvia/core/vector-store';
import { QdrantVectorClient } from '@anvia/qdrant';
const storeClient = new QdrantVectorClient({
    url: process.env.QDRANT_URL!,
    apiKey: process.env.QDRANT_API_KEY!,
});
const store = storeClient.vectorStore({
    collectionName: 'support_docs',
    dimensions: 1536,
    metric: "cosine"
});
await store.ensure();
await store.upsert({
    documents: documents
});
const results = await retrieveDocuments({ store, model: embeddings, query, topK: 5 });
const stored = await store.get({ documentIds: ['support-1'] });
const page = await store.inspect({ limit: 50 });

Pass either an existing client or official Qdrant client parameters such as url and apiKey. Use ensure() when this process may create a missing collection and validate() when infrastructure owns provisioning.

For hybrid search, create the store with mode: 'hybrid'; ingested documents must contain aligned dense and sparse embeddings.

Next

• Collections and indexing
• Search and filters
• Production