@orka-js/qdrant

Qdrant vector database adapter for OrkaJS — high-performance open-source vector search.

Installation

npm install @orka-js/qdrant

Start Qdrant locally with Docker:

docker run -p 6333:6333 qdrant/qdrant

Quick Start

import { QdrantAdapter } from '@orka-js/qdrant'
import { Orka } from '@orka-js/core'

const vectorDB = new QdrantAdapter({
  url: 'http://localhost:6333',
})

const orka = new Orka({ llm, vectorDB })

await orka.knowledge.ingest([
  { id: 'doc-1', content: 'Hello world' }
])
const results = await orka.knowledge.query('hello', 5)

Direct Usage

const adapter = new QdrantAdapter({ url: 'http://localhost:6333' })

await adapter.createCollection('documents', 1536, 'cosine')
await adapter.upsert('documents', [
  { id: 'doc-1', vector: [...], metadata: { text: 'Hello world' } }
])
const results = await adapter.search('documents', queryVector, 5)
await adapter.delete('documents', ['doc-1'])

Configuration

| Option | Type | Default | Description | |--------|------|---------|-------------| | url | string | — | Qdrant server URL (required) | | apiKey | string | — | API key for Qdrant Cloud | | collectionName | string | — | Default collection name |

Distance Metrics

| Value | Description | |-------|-------------| | cosine | Cosine similarity (recommended for normalized embeddings) | | euclidean | L2 distance | | dotProduct | Dot product (for un-normalized embeddings) |

API

QdrantAdapter

adapter.createCollection(name, dimension, distance?)  // Promise<void>
adapter.deleteCollection(name)                        // Promise<void>
adapter.upsert(collection, docs)                      // Promise<void>
adapter.search(collection, vector, topK, filter?)     // Promise<SearchResult[]>
adapter.delete(collection, ids)                       // Promise<void>

Implements: VectorDBAdapter

Related Packages

• @orka-js/core — VectorDBAdapter interface
• @orka-js/pinecone — Managed cloud alternative
• @orka-js/memory — In-memory adapter for dev/testing
• orkajs — Full bundle