# Qdrant

> A Rust vector search engine with rich payload filtering for semantic search

- Category: Retrieval & Knowledge (RAG) / Vector Databases
- Language: Rust
- Licence: Apache-2.0
- Repository: https://github.com/qdrant/qdrant
- Website: https://qdrant.tech
- Canonical page: https://ai-tldr.dev/tools/qdrant/

## In short

Qdrant (read: "quadrant") is an open-source vector similarity search engine and vector database. It stores points - vectors paired with an extra payload - and lets you search them by similarity, with filtering on the payload data. It sits in the Vector Databases category of the Retrieval & Knowledge (RAG) tools and is open-source, written in Rust and licensed Apache-2.0.

## Overview

Qdrant (read: "quadrant") is an open-source vector similarity search engine and vector database. It stores points - vectors paired with an extra payload - and lets you search them by similarity, with filtering on the payload data. It exposes a service with an API so you can store, search, and manage these vectors from your application.

It is written in Rust, which helps it stay fast and reliable under high load. Qdrant is built for teams that turn embeddings or neural-network encoders into real applications for matching, searching, and recommending. Its extended filtering support also makes it useful for faceted search and other semantic matching tasks.

As a vector database, Qdrant sits between your embedding model and your app. You run it as a server (locally via Docker or self-hosted), or use the fully managed Qdrant Cloud, which includes a free tier. A lightweight option, Qdrant Edge, can run inside your application process for offline or low-latency use.

## What it does

- Vector similarity search with an extended filtering layer over the payload attached to each point
- Written in Rust for speed and stability under high request load
- Official client libraries for Python, JavaScript/TypeScript, Go, Rust, .NET/C#, and Java
- Production-ready service with a convenient API to store, search, and manage points
- Qdrant Edge runs inside your application process for offline and low-latency use on edge devices
- Available self-hosted via Docker or as the fully managed Qdrant Cloud with a free tier

## Getting started

Run Qdrant locally with Docker, then connect to it with one of the official clients such as Python.

1. Start the Qdrant server

   Run the container to start Qdrant locally on port 6333. Note this starts an insecure deployment without authentication, so secure your instance before production.

```bash
docker run -p 6333:6333 qdrant/qdrant
```

2. Connect with the Python client

   Point the client at your running server to begin storing and searching vectors.

```python
from qdrant_client import QdrantClient

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

3. Read the docs before production

   Before deploying to production, follow the official installation and security guides at qdrant.tech, and check the Quick Start guide for a full walkthrough.

## When to use it

- Build a semantic text search engine that finds meaning rather than matching keywords
- Power recommendation systems by matching embeddings from a neural-network encoder
- Run similar-image search, such as visual food discovery, without relying on text descriptions
- Add faceted or filtered semantic search by combining vector similarity with payload filters