Qdrant

Founded by Andre Zayarni in 2021

Store, search, and manage high-dimensional vector embeddings for machine learning applications

Pricing model

Free Tier, Paid

Rating

★ People love it

Time to value

Quick Setup (< 1 hour)

You can use Qdrant to store and search vector embeddings for retrieval augmented generation, semantic search, and recommendation engines. It is an open-source vector database written in Rust that handles dense and sparse vectors, along with rich JSON payload filtering. You can run similarity searches with low latency while applying complex metadata filters during search execution. Qdrant also supports multivector representations, quantization to save memory, and real-time indexing so newly added vectors are immediately searchable through REST or gRPC APIs.

What Qdrant does

Create and manage vector collections with custom distance metricsUpload document embeddings alongside JSON metadata payloadsRun similarity queries with filter conditions on metadata attributesConfigure quantization settings to reduce cluster memory footprintQuery collections via REST or gRPC client librariesMonitor database performance metrics using Prometheus and GrafanaRerank retrieval results using late interaction modelsNative hybrid search combining dense and sparse vectorsSingle-stage JSON metadata filtering during HNSW traversalScalar, product, and binary quantization to reduce memory usageReal-time vector indexing without rebuilding entire collectionsMultivector support for storing multiple vectors per objectBuilt-in web visual interface for querying and inspecting collectionsAPI access via REST, gRPC, and official programming SDKs