Ecosystem — Qdrant Course

Official Client Libraries

Qdrant provides official clients for six languages plus direct REST/gRPC access from any language.

Python (qdrant-client) ▼

The most feature-complete client. Supports sync/async operations, local mode (in-memory or persistent, no server needed), automatic batching, and full type hints.

python

from qdrant_client import QdrantClient
client = QdrantClient(":memory:")  # local mode
client = QdrantClient(url="http://localhost:6333")  # remote

JavaScript/TypeScript ▼

REST-based client for Node.js and browser. Full TypeScript type definitions. Package: @qdrant/js-client-rest.

Rust, Go, .NET, Java ▼

Native Rust client uses gRPC for maximum performance. Go and Java clients also use gRPC. The .NET client (Qdrant.Client) supports both REST and gRPC.

Framework Integrations

LangChain ▼

First-class vector store via the langchain-qdrant partner package. Supports dense, sparse, and hybrid retrieval.

python

from langchain_qdrant import QdrantVectorStore
vs = QdrantVectorStore.from_documents(
    docs, embedding=OpenAIEmbeddings(),
    url="http://localhost:6333",
    collection_name="langchain_docs",
)
retriever = vs.as_retriever(search_kwargs={"k": 5})

LlamaIndex ▼

Vector store backend for LlamaIndex. Handles document chunking, embedding, and storage. LlamaIndex manages ingestion while Qdrant provides the search layer.

Haystack, Semantic Kernel, AutoGen ▼

Deepset Haystack includes QdrantDocumentStore. Microsoft Semantic Kernel uses Qdrant as a memory backend. Multi-agent frameworks (AutoGen, CrewAI) use Qdrant for long-term agent memory.

Deployment Options

Open Source (Self-Hosted)

Run via Docker or compile from source. Full feature set, unlimited scale, no license fees. You manage infrastructure, backups, and upgrades.

Qdrant Cloud (Managed)

Fully managed on AWS, GCP, or Azure. Includes scaling, backups, monitoring, and a free tier (1GB). Production deployments get SLA, SSO, and RBAC.

Hybrid Cloud & Qdrant Edge

Hybrid Cloud: your infrastructure, Qdrant management plane. Edge: in-process version sharing the server's storage format for low-latency offline-capable search on devices.

Common Integration Patterns

Qdrant + LangChain/LlamaIndex for RAG

Documents are chunked, embedded, and stored in Qdrant. At query time, the framework embeds the question, retrieves chunks, and passes them to the LLM. Payload filtering scopes by source, date, or access level.

Qdrant + Fastembed for Self-Contained Search

Fastembed generates embeddings locally without external API calls -- ideal for privacy-sensitive or offline deployments.

Qdrant + PostgreSQL for Hybrid Storage

Store structured data in PostgreSQL, vector embeddings in Qdrant. Application logic queries both: PostgreSQL for relational queries, Qdrant for similarity search. Join results in the application layer.

Ecosystem & Integrations