Qdrant

Qdrant is presented as a high-performance vector search engine for building AI retrieval. The page highlights a Rust-based engine with SIMD and a custom storage engine (Gridstore), real-time indexing of new vectors, memory-efficient storage, and quantization that the page says can reduce memory usage by up to 64x. It offers hybrid dense and sparse search (citing BM25, SPLADE++, and miniCOIL), metadata filtering, multivector support, and reranking options. Developer access is via REST, gRPC, and official clients such as Python and JavaScript, plus a built-in web UI. Deployment options include Qdrant Cloud on AWS, GCP, or Azure, Hybrid Cloud, Private Cloud, and Qdrant Edge (Beta). The page also mentions Qdrant Cloud Inference for generating text and image embeddings, and lists SOC 2, GDPR-aligned options, SSO, RBAC, and backups. Use cases named on the page include RAG, AI agents, semantic search, recommendation systems, and data analysis and anomaly detection. Pricing is not detailed on this page beyond a link to a pricing section and a 'Start Free' call to action.

Key features

• Vector similarity search written in Rust with SIMD and custom Gridstore storage engine
• Real-time indexing where new vectors are searchable once added
• Asymmetric, scalar, and binary quantization reducing memory usage by up to 64x
• Hybrid dense and sparse search supporting BM25, SPLADE++, and miniCOIL
• Expansive metadata filters including nested, text, geo, and has_vector
• Built-in multivector support for multiple vectors per object
• Reranking with score boosting, late interaction models such as ColBERT, and Maximum Marginal Relevance
• Developer APIs via REST, gRPC, and official clients such as Python and JavaScript
• Built-in web UI for exploring collections and testing queries
• Native cloud inference for generating text and image embeddings in Qdrant Cloud
• Deployment options including Qdrant Cloud, Hybrid Cloud, Private Cloud, and Edge (Beta)
• Enterprise tooling such as SSO, multitenancy with granular RBAC, backups with point-in-time restore, and vector-scoped API keys

Best suited for

• Retrieval-augmented generation (RAG) and GenAI applications
• AI agents needing fast similarity search and memory
• Semantic search
• Recommendation systems
• Data analysis and anomaly detection
• Teams deploying vector search on cloud, hybrid, on-prem, or edge environments

Pricing

Not stated on the website