spb/tools-spboucher-ai
Public
1{2 "aliases": [],3 "workflows": [4 "Vector store for a RAG pipeline built with LangChain or LlamaIndex",5 "Hybrid keyword and semantic search over document collections",6 "Recommendation and deduplication with payload filters",7 "Single-binary local deployment on an Apple Silicon workstation"8 ],9 "advantages": [10 "Apache 2.0 license and self-hostable single binary",11 "Filtering integrated with the vector index",12 "Memory reduction via quantization",13 "Managed cloud with free tier"14 ],15 "limitations": [16 "No embedding generation on the server",17 "No SQL or relational joins",18 "Payload filters need indexes to avoid slow scans",19 "Cluster operations add complexity compared with pgvector inside Postgres"20 ],21 "prerequisites": [22 "Docker or a native binary, or a Qdrant Cloud account",23 "An embedding model or API to produce vectors"24 ],25 "audiences": [26 "Developers building RAG and semantic search",27 "ML engineers running retrieval experiments"28 ],29 "alternatives": [30 "pgvector",31 "weaviate",32 "milvus",33 "chroma",34 "pinecone",35 "elasticsearch"36 ],37 "tags": [38 "vector-search",39 "embeddings",40 "rag",41 "rust",42 "self-hosted",43 "open-source",44 "api"45 ],46 "use_cases": [47 "build-agents",48 "semantic-search"49 ],50 "operating_systems": [51 "linux",52 "macos",53 "windows"54 ],55 "deployment_models": [56 "on-premises",57 "managed-cloud",58 "hybrid"59 ],60 "integrations": [61 "LangChain",62 "LlamaIndex",63 "Haystack",64 "FastEmbed",65 "OpenAI embeddings",66 "Cohere",67 "Voyage AI"68 ],69 "data_formats": [70 "JSON payloads",71 "Snapshot archives"72 ],73 "languages": [74 "Rust",75 "Python",76 "TypeScript",77 "Go",78 "Java",79 "C#"80 ],81 "self_hostable": true,82 "discovery_method": "editorial-research",83 "depth": "deep",84 "editorial_status": "reviewed",85 "visibility": "public",86 "featured": false,87 "author": "batch-D",88 "slug": "qdrant",89 "name": "Qdrant",90 "vendor": {91 "name": "Qdrant Solutions GmbH",92 "slug": "qdrant-solutions",93 "website": "https://qdrant.tech/"94 },95 "homepage_url": "https://qdrant.tech/",96 "docs_url": "https://qdrant.tech/documentation/",97 "repo_url": "https://github.com/qdrant/qdrant",98 "license_url": "https://github.com/qdrant/qdrant/blob/master/LICENSE",99 "pricing_url": "https://qdrant.tech/pricing/",100 "pricing_checked_at": "2026-11-02",101 "product_type": "database",102 "product_status": "active",103 "summary": "Open-source vector search engine written in Rust with filterable HNSW indexes, sparse and multi-vector support, REST/gRPC APIs and a managed cloud",104 "description": "Qdrant is an open-source vector database and similarity search engine written in Rust by Qdrant Solutions GmbH and released under the Apache 2.0 license. It stores points consisting of one or more vectors plus a JSON payload, builds HNSW indexes over them, and answers nearest-neighbor queries that can be constrained by payload filters evaluated during the graph search rather than after it. Dense, sparse and multi-vector (late-interaction) representations can coexist in one collection, which supports hybrid lexical-plus-semantic retrieval and reranking pipelines.\n\nThe server exposes REST and gRPC APIs with official clients for Python, JavaScript/TypeScript, Rust, Go, Java and .NET, and it integrates with LangChain, LlamaIndex, Haystack and most embedding providers. Scalar, product and binary quantization reduce memory, and on-disk payload and vector storage allow collections larger than RAM. Distributed mode shards and replicates collections across nodes with Raft-based consensus; snapshots cover backup and migration. Qdrant Cloud offers managed clusters with a free tier, and a hybrid option runs the data plane in the customer's own infrastructure.\n\nQdrant is a common choice for retrieval-augmented generation, semantic search, recommendation and deduplication where metadata filtering matters. It does not compute embeddings itself (apart from the optional FastEmbed helper in client libraries), has no SQL interface or joins, and filter-heavy queries need payload indexes to stay fast. Operating a multi-node cluster requires attention to shard placement and memory, and the single-binary deployment on macOS or Linux is often sufficient for research-scale collections.",105 "capabilities": [106 "HNSW approximate nearest-neighbor search with in-graph payload filtering",107 "Dense, sparse and multi-vector named vectors per point",108 "Scalar, product and binary quantization",109 "REST and gRPC APIs with clients in six languages",110 "Distributed sharding and replication with Raft consensus",111 "Snapshots and on-disk storage for collections larger than memory",112 "Hybrid search and server-side fusion/reranking queries"113 ],114 "comparison_notes": "pgvector keeps vectors inside PostgreSQL with SQL and transactions but without a dedicated distributed index service. Weaviate bundles vectorizer modules and a GraphQL API; Milvus targets very large distributed deployments with many index types; Chroma emphasizes a minimal developer API for prototypes. Pinecone is proprietary and cloud-only. Qdrant's distinguishing traits are its Rust implementation, filterable HNSW and quantization options under a permissive license.",115 "categories": [116 {117 "slug": "vector-databases",118 "primary": true119 },120 {121 "slug": "embeddings-vector-search"122 }123 ],124 "execution_modes": [125 "self-hosted",126 "cloud",127 "local"128 ],129 "interfaces": [130 "api",131 "sdk",132 "web"133 ],134 "source_model": "open-source",135 "license_spdx": "Apache-2.0",136 "pricing_model": [137 "open-source",138 "freemium",139 "usage-based",140 "enterprise"141 ],142 "sources": [143 {144 "url": "https://qdrant.tech/",145 "type": "official_site",146 "supports": [147 "name",148 "summary"149 ],150 "checked_at": "2026-11-02",151 "title": "Qdrant – Vector Database"152 },153 {154 "url": "https://qdrant.tech/documentation/",155 "type": "official_docs",156 "supports": [157 "capabilities",158 "interfaces",159 "languages"160 ],161 "checked_at": "2026-11-02",162 "title": "Qdrant Documentation"163 },164 {165 "url": "https://github.com/qdrant/qdrant/blob/master/LICENSE",166 "type": "license_page",167 "supports": [168 "license_spdx",169 "source_model"170 ],171 "checked_at": "2026-11-02",172 "title": "Qdrant LICENSE (Apache 2.0)"173 },174 {175 "url": "https://qdrant.tech/pricing/",176 "type": "pricing_page",177 "supports": [178 "pricing_model"179 ],180 "checked_at": "2026-11-02",181 "title": "Qdrant Cloud pricing"182 },183 {184 "url": "https://github.com/qdrant/qdrant",185 "type": "repository",186 "supports": [187 "repo_url"188 ],189 "checked_at": "2026-11-02",190 "title": "qdrant/qdrant"191 }192 ],193 "confidence": "high"194}195