tools
Qdrant
Qdrant is an open-source vector search engine and managed cloud service for semantic search, RAG, recommendations, and AI agents.
In inglese
Qdrant stores vectors with metadata and supports similarity search, metadata filtering, dense and sparse hybrid search, multivector retrieval, and reranking. It provides REST and gRPC APIs, official client libraries, a Web UI, and managed cloud deployment.
Developers use it for semantic search, retrieval-augmented generation, recommendation systems, anomaly detection, and agent memory. Qdrant can run locally, in hybrid or private environments, or in Qdrant Cloud. Cloud pricing is usage-based for production clusters, and paid embedding models may add token charges.
Features
- Dense and sparse hybrid search with BM25, SPLADE++, and miniCOIL
- Metadata filtering with nested, text, geo, and has_vector conditions
- Multivector retrieval for text, image, audio, and other multimodal data
- Reranking with score boosting, ColBERT, and Maximum Marginal Relevance
- Real-time indexing without rebuilding the entire index
- Scalar, binary, asymmetric, and TurboQuant quantization
- REST and gRPC APIs with Python, TypeScript, Rust, Go, Java, and .NET clients
- Open-source Apache 2.0 engine available for local deployment
Use cases
- Build retrieval-augmented generation systems
- Search documents by meaning rather than exact keywords
- Recommend products, content, or experiences using vector similarity
- Give AI agents persistent memory and contextual retrieval
- Detect anomalies by finding unusual vector patterns
- Run filtered hybrid search across large enterprise collections
Pros
Cons
Latest updates
- v1.19.0 (v1.19.0)
Added 4-bit TurboQuant storage, per-component memory controls, keyword prefix filters, and per-query IDF for sparse search.
- v1.18.0 (v1.18.0)
- v1.17.0 (v1.17.0)
Added relevance feedback, optimization and telemetry APIs, update queue and controls, audit logging, weighted RRF, and upsert modes.
Capabilities
- Command line — “API, Terraform, Pulumi, CLI” source
- Self-hosted — “Open-source DNA with enterprise-grade security and flexibility — run on-prem, hybrid, edge, or move seamlessly to Qdrant Cloud.” source
- API — “Start with a single API call — scale to advanced control over HNSW, hybrid fusion, reranking, and multi-vector retrieval, all via REST, gRPC, or official clients (Python, JavaScript, etc.).” source
- Official SDKs — “official clients (Python, JavaScript, etc.).” source
- Runs models for you — “Generate text and image embeddings and run vector search in Qdrant Cloud — no separate pipeline or infrastructure needed.” source
- Search over your data — “Deliver context-rich answers with hybrid dense – sparse retrieval, metadata filters, and fresh updates.” source
Security
- SOC 2 Type II — “Qdrant holds SOC 2 Type 2 and HIPAA certifications.” source
- SOC 2 — “Qdrant holds SOC 2 Type 2 and HIPAA certifications.” source
- GDPR — “For deployments subject to GDPR, Qdrant provides a Data Processing Agreement covering data protection and privacy commitments.” source
- HIPAA — “Qdrant is HIPAA certified and Business Associate Agreement is available for Qdrant Managed Cloud.” source
- Data kept in the EU — “Data in Qdrant clusters stored only in the cluster's deployment region.” source
Pricing
- Starting price
- Free
- Prices checked
- 2026-09-25
Free Tier
Free
- Free forever
- Single Node Cluster
- 0.5 vCPU / 1GB RAM / 4 GB Disk
- Free Cloud Inference With Selected Models
Standard Tier
- from 25$
- Usage-based pricing
- Dedicated Resources
- Flexible Vertical and Horizontal Scaling
- Highly Available Setups
- Backup & Disaster Recovery
- 99.5% Uptime SLA