Skip to main content

Vector Search Engine for Next-Generation AI

Build powerful semantic search, recommendation systems, and RAG applications with Qdrant — a high-performance vector database designed for production workloads.

Get StartedAPI Reference

Quick start

Get Qdrant running in minutes

1

Run Qdrant with Docker

Start a Qdrant instance locally using Docker:
docker run -p 6333:6333 qdrant/qdrant
This starts an insecure deployment for development. For production, see our security guide.
2

Install a client library

Install the Qdrant client for your language:
pip install qdrant-client
3

Create your first collection

Connect to Qdrant and create a collection to store vectors:
from qdrant_client import QdrantClient
from qdrant_client.models import Distance, VectorParams

client = QdrantClient("http://localhost:6333")

client.create_collection(
    collection_name="my_collection",
    vectors_config=VectorParams(size=384, distance=Distance.COSINE)
)

{
  "result": true,
  "status": "ok",
  "time": 0.031
}
4

Insert and search vectors

Add vectors with payloads and perform similarity search:
from qdrant_client.models import PointStruct

# Insert points
client.upsert(
    collection_name="my_collection",
    points=[
        PointStruct(id=1, vector=[0.05, 0.61, 0.76, ...], payload={"city": "Berlin"}),
        PointStruct(id=2, vector=[0.19, 0.81, 0.75, ...], payload={"city": "London"}),
    ]
)

# Search
results = client.query_points(
    collection_name="my_collection",
    query=[0.2, 0.1, 0.9, ...],
    limit=3
)
Learn more in our quickstart guide.

Explore by topic

Deep dive into Qdrant’s capabilities

Core Concepts

Understand collections, points, vectors, and payloads

Working with Data

Insert, search, filter, and manage your vector data

Hybrid Search

Combine dense and sparse vectors for better results

Quantization

Reduce memory usage by up to 97% with quantization

Distributed Deployment

Scale horizontally with sharding and replication

Performance Tuning

Optimize search speed and resource usage

Choose your interface

Access Qdrant through REST, gRPC, or client libraries

REST API

HTTP API with OpenAPI specification for easy integration

gRPC API

High-performance gRPC interface for production workloads

Python Client

Full-featured client with async support and type hints

JavaScript Client

TypeScript-ready client for Node.js and browsers

Rust Client

Native Rust client with zero-copy deserialization

More Languages

Go, .NET, Java, and community-maintained clients

Ready to build with Qdrant?

Start with our quickstart guide or explore the full API reference to integrate vector search into your application.

Get Started NowStar on GitHub