Technology RadarTechnology Radar
This item was not updated in last three versions of the Radar. Should it have appeared in one of the more recent editions, there is a good chance it remains pertinent. However, if the item dates back further, its relevance may have diminished and our current evaluation could vary. Regrettably, our capacity to consistently revisit items from past Radar editions is limited.
Trial

Qdrant is an open-source vector database written in Rust — the strongest self-hosted option for teams that want Pinecone-class performance without the SaaS dependency.

Buy vs Build

Qdrant is primarily build (self-hosted, open-source), but Qdrant Cloud is a managed option if you want to buy your way out of operations. The open-source version is the most widely-used path.

Why It's in Trial

Qdrant is the recommended choice when you need self-hosted vector search with production-grade reliability. Compared to alternatives:

  • vs pgvector: Faster at scale (>10M vectors), better filtering performance, purpose-built for vector search
  • vs Pinecone: Self-hosted (no data egress to third party), no per-query costs, full control
  • vs Weaviate: Simpler setup, better raw performance benchmarks, Rust-based stability

What makes Qdrant stand out:

  • Filtering performance: The ACORN algorithm (2025) makes filtered HNSW competitive even with very narrow filters — you can combine metadata filtering with vector search without sacrificing accuracy
  • Payload indexing: Efficient structured data filtering alongside vector similarity
  • SOC 2 Type II (Qdrant Cloud)
  • gRPC + REST: Dual API support; gRPC for performance-sensitive paths
  • Horizontal scaling: Native distributed mode with sharding and replication

When to Choose Qdrant

  • You need self-hosted for data sovereignty or compliance reasons
  • You want fine-grained control over infrastructure and costs
  • You need rich metadata filtering alongside vector search
  • You're building in a regulated environment and can't use SaaS

Getting Started with Docker

docker pull qdrant/qdrant
docker run -p 6333:6333 -v ./qdrant_storage:/qdrant/storage qdrant/qdrant

Then interact via the Python SDK:

from qdrant_client import QdrantClient

client = QdrantClient("localhost", port=6333)

Key Characteristics

Property Value
Written in Rust
License Apache 2.0
Managed option Qdrant Cloud (SOC 2 Type II)
API REST + gRPC
Provider Qdrant Inc.
Website qdrant.tech
GitHub qdrant/qdrant

Further Reading