[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-qdrant-qdrant-memory-usage-optimization":3,"mdc--bz9w4g-key":45,"related-org-qdrant-qdrant-memory-usage-optimization":363,"related-repo-qdrant-qdrant-memory-usage-optimization":515},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":20,"repoUrl":21,"updatedAt":22,"license":23,"forks":24,"topics":25,"repo":40,"sourceUrl":43,"mdContent":44},"qdrant-memory-usage-optimization","optimize Qdrant memory usage","Diagnoses and reduces Qdrant memory usage. Use when someone reports 'memory too high', 'RAM keeps growing', 'node crashed', 'out of memory', 'memory leak', or asks 'why is memory usage so high?', 'how to reduce RAM?'. Also use when memory doesn't match calculations, quantization didn't help, or nodes crash during recovery.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},"qdrant","Qdrant","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fqdrant.png",[12,16,17],{"name":13,"slug":14,"type":15},"Performance","performance","tag",{"name":9,"slug":8,"type":15},{"name":18,"slug":19,"type":15},"Debugging","debugging",197,"https:\u002F\u002Fgithub.com\u002Fqdrant\u002Fskills","2026-07-16T06:00:15.357243",null,23,[26,27,28,29,30,31,32,33,34,14,8,35,36,37,38,39],"agent-skills","ai-agents","claude-code","codex","cursor","embeddings","hybrid-search","monitoring","multitenancy","quantization","scaling","search-quality","vector-database","vector-search",{"repoUrl":21,"stars":20,"forks":24,"topics":41,"description":42},[26,27,28,29,30,31,32,33,34,14,8,35,36,37,38,39],"Agent skills for Qdrant vector search: scaling, performance optimization, search quality, monitoring, deployment, model migration, version upgrades, and SDK usage across Python, TypeScript, Rust, Go, .NET, Java","https:\u002F\u002Fgithub.com\u002Fqdrant\u002Fskills\u002Ftree\u002FHEAD\u002Fskills\u002Fqdrant-performance-optimization\u002Fmemory-usage-optimization","---\nname: qdrant-memory-usage-optimization\ndescription: \"Diagnoses and reduces Qdrant memory usage. Use when someone reports 'memory too high', 'RAM keeps growing', 'node crashed', 'out of memory', 'memory leak', or asks 'why is memory usage so high?', 'how to reduce RAM?'. Also use when memory doesn't match calculations, quantization didn't help, or nodes crash during recovery.\"\n---\n\n# Understanding memory usage\n\nQdrant operates with two types of memory:\n\n- Resident memory (aka RSSAnon) - memory used for internal data structures like the ID tracker, plus components that must stay in RAM, such as quantized vectors when `always_ram=true` and payload indexes.\n\n- OS page cache - memory used for caching disk reads, which can be released when needed. Original vectors are normally stored in page cache, so the service won't crash if RAM is full, but performance may degrade.\n\nIt is normal for the OS page cache to occupy all available RAM, but if resident memory is above 80% of total RAM, it is a sign of a problem.\n\n## Memory usage monitoring\n\n- Qdrant exposes memory usage through the `\u002Fmetrics` endpoint. See [Monitoring docs](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fops-monitoring\u002Fmonitoring\u002F).\n\n\u003C!-- ToDo: Talk about memory usage of each components once API is available -->\n\n\n## How much memory is needed for Qdrant?\n\nOptimal memory usage depends on the use case.\n\n- For regular search scenarios, general guidelines are provided in the [Capacity planning docs](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fcapacity-planning\u002F).\n\nFor a detailed breakdown of memory usage at large scale, see [Large scale memory usage example](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Ftutorials-operations\u002Flarge-scale-search\u002F?s=memory-usage).\n\nPayload indexes and HNSW graph also require memory, along with vectors themselves, so it's important to consider them in calculations.\n\nAdditionally, Qdrant requires some extra memory for optimizations. During optimization, optimized segments are fully loaded into RAM, so it is important to leave enough headroom.\nThe larger `max_segment_size` is, the more headroom is needed.\n\n\n### When to put HNSW index on disk\n\nPutting frequently used components (such as HNSW index) on disk might cause significant performance degradation.\nThere are some scenarios, however, when it can be a good option:\n\n- Deployments with low latency disks - local NVMe or similar.\n- Multi-tenant deployments, where only a subset of tenants is frequently accessed, so that only a fraction of data & index is loaded in RAM at a time.\n- For deployments with [inline storage](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fops-optimization\u002Foptimize\u002F?s=inline-storage-in-hnsw-index) enabled.\n\n\n## How to minimize memory footprint\n\nThe main challenge is to put on disk those parts of data, which are rarely accessed.\nHere are the main techniques to achieve that:\n\n- Use quantization to store only compressed vectors in RAM [Quantization docs](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fquantization\u002F)\n\n- Use float16 or int8 datatypes to reduce memory usage of vectors by 2x or 4x respectively, with some tradeoff in precision. Read more about vector datatypes in [documentation](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fvectors\u002F?s=datatypes)\n\n- Leverage Matryoshka Representation Learning (MRL) to store only small vectors in RAM while keeping large vectors on disk. Examples of how to use MRL with Qdrant Cloud inference: [MRL docs](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Finference\u002Fmatryoshka-models\u002F?s=reduce-vector-dimensionality-with-matryoshka-models)\n\n- For multi-tenant deployments with small tenants, vectors might be stored on disk because the same tenant's data is stored together [Multitenancy docs](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fmultitenancy\u002F?s=calibrate-performance)\n\n- For deployments with fast local storage and relatively low requirements for search throughput, it may be possible to store all components of vector store on disk. Read more about the performance implications of on-disk storage in [the article](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Farticles\u002Fmemory-consumption\u002F)\n\n- For low RAM environments, consider `async_scorer` config, which enables support of `io_uring` for parallel disk access, which can significantly improve performance of on-disk storage. Read more about `async_scorer` in [the article](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Farticles\u002Fio_uring\u002F) (only available on Linux with kernel 5.11+)\n\n- Consider storing Sparse Vectors and text payload on disk, as they are usually more disk-friendly than dense vectors.\n- Configure payload indexes to be stored on disk [docs](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Findexing\u002F?s=on-disk-payload-index)\n- Configure sparse vectors to be stored on disk [docs](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Findexing\u002F?s=sparse-vector-index)\n\n",{"data":46,"body":47},{"name":4,"description":6},{"type":48,"children":49},"root",[50,59,65,89,94,101,128,134,139,155,168,173,186,193,198,225,231,236],{"type":51,"tag":52,"props":53,"children":55},"element","h1",{"id":54},"understanding-memory-usage",[56],{"type":57,"value":58},"text","Understanding memory usage",{"type":51,"tag":60,"props":61,"children":62},"p",{},[63],{"type":57,"value":64},"Qdrant operates with two types of memory:",{"type":51,"tag":66,"props":67,"children":68},"ul",{},[69,84],{"type":51,"tag":70,"props":71,"children":72},"li",{},[73,75,82],{"type":57,"value":74},"Resident memory (aka RSSAnon) - memory used for internal data structures like the ID tracker, plus components that must stay in RAM, such as quantized vectors when ",{"type":51,"tag":76,"props":77,"children":79},"code",{"className":78},[],[80],{"type":57,"value":81},"always_ram=true",{"type":57,"value":83}," and payload indexes.",{"type":51,"tag":70,"props":85,"children":86},{},[87],{"type":57,"value":88},"OS page cache - memory used for caching disk reads, which can be released when needed. Original vectors are normally stored in page cache, so the service won't crash if RAM is full, but performance may degrade.",{"type":51,"tag":60,"props":90,"children":91},{},[92],{"type":57,"value":93},"It is normal for the OS page cache to occupy all available RAM, but if resident memory is above 80% of total RAM, it is a sign of a problem.",{"type":51,"tag":95,"props":96,"children":98},"h2",{"id":97},"memory-usage-monitoring",[99],{"type":57,"value":100},"Memory usage monitoring",{"type":51,"tag":66,"props":102,"children":103},{},[104],{"type":51,"tag":70,"props":105,"children":106},{},[107,109,115,117,126],{"type":57,"value":108},"Qdrant exposes memory usage through the ",{"type":51,"tag":76,"props":110,"children":112},{"className":111},[],[113],{"type":57,"value":114},"\u002Fmetrics",{"type":57,"value":116}," endpoint. See ",{"type":51,"tag":118,"props":119,"children":123},"a",{"href":120,"rel":121},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fops-monitoring\u002Fmonitoring\u002F",[122],"nofollow",[124],{"type":57,"value":125},"Monitoring docs",{"type":57,"value":127},".",{"type":51,"tag":95,"props":129,"children":131},{"id":130},"how-much-memory-is-needed-for-qdrant",[132],{"type":57,"value":133},"How much memory is needed for Qdrant?",{"type":51,"tag":60,"props":135,"children":136},{},[137],{"type":57,"value":138},"Optimal memory usage depends on the use case.",{"type":51,"tag":66,"props":140,"children":141},{},[142],{"type":51,"tag":70,"props":143,"children":144},{},[145,147,154],{"type":57,"value":146},"For regular search scenarios, general guidelines are provided in the ",{"type":51,"tag":118,"props":148,"children":151},{"href":149,"rel":150},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fcapacity-planning\u002F",[122],[152],{"type":57,"value":153},"Capacity planning docs",{"type":57,"value":127},{"type":51,"tag":60,"props":156,"children":157},{},[158,160,167],{"type":57,"value":159},"For a detailed breakdown of memory usage at large scale, see ",{"type":51,"tag":118,"props":161,"children":164},{"href":162,"rel":163},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Ftutorials-operations\u002Flarge-scale-search\u002F?s=memory-usage",[122],[165],{"type":57,"value":166},"Large scale memory usage example",{"type":57,"value":127},{"type":51,"tag":60,"props":169,"children":170},{},[171],{"type":57,"value":172},"Payload indexes and HNSW graph also require memory, along with vectors themselves, so it's important to consider them in calculations.",{"type":51,"tag":60,"props":174,"children":175},{},[176,178,184],{"type":57,"value":177},"Additionally, Qdrant requires some extra memory for optimizations. During optimization, optimized segments are fully loaded into RAM, so it is important to leave enough headroom.\nThe larger ",{"type":51,"tag":76,"props":179,"children":181},{"className":180},[],[182],{"type":57,"value":183},"max_segment_size",{"type":57,"value":185}," is, the more headroom is needed.",{"type":51,"tag":187,"props":188,"children":190},"h3",{"id":189},"when-to-put-hnsw-index-on-disk",[191],{"type":57,"value":192},"When to put HNSW index on disk",{"type":51,"tag":60,"props":194,"children":195},{},[196],{"type":57,"value":197},"Putting frequently used components (such as HNSW index) on disk might cause significant performance degradation.\nThere are some scenarios, however, when it can be a good option:",{"type":51,"tag":66,"props":199,"children":200},{},[201,206,211],{"type":51,"tag":70,"props":202,"children":203},{},[204],{"type":57,"value":205},"Deployments with low latency disks - local NVMe or similar.",{"type":51,"tag":70,"props":207,"children":208},{},[209],{"type":57,"value":210},"Multi-tenant deployments, where only a subset of tenants is frequently accessed, so that only a fraction of data & index is loaded in RAM at a time.",{"type":51,"tag":70,"props":212,"children":213},{},[214,216,223],{"type":57,"value":215},"For deployments with ",{"type":51,"tag":118,"props":217,"children":220},{"href":218,"rel":219},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fops-optimization\u002Foptimize\u002F?s=inline-storage-in-hnsw-index",[122],[221],{"type":57,"value":222},"inline storage",{"type":57,"value":224}," enabled.",{"type":51,"tag":95,"props":226,"children":228},{"id":227},"how-to-minimize-memory-footprint",[229],{"type":57,"value":230},"How to minimize memory footprint",{"type":51,"tag":60,"props":232,"children":233},{},[234],{"type":57,"value":235},"The main challenge is to put on disk those parts of data, which are rarely accessed.\nHere are the main techniques to achieve that:",{"type":51,"tag":66,"props":237,"children":238},{},[239,251,263,275,287,299,335,340,352],{"type":51,"tag":70,"props":240,"children":241},{},[242,244],{"type":57,"value":243},"Use quantization to store only compressed vectors in RAM ",{"type":51,"tag":118,"props":245,"children":248},{"href":246,"rel":247},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fquantization\u002F",[122],[249],{"type":57,"value":250},"Quantization docs",{"type":51,"tag":70,"props":252,"children":253},{},[254,256],{"type":57,"value":255},"Use float16 or int8 datatypes to reduce memory usage of vectors by 2x or 4x respectively, with some tradeoff in precision. Read more about vector datatypes in ",{"type":51,"tag":118,"props":257,"children":260},{"href":258,"rel":259},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fvectors\u002F?s=datatypes",[122],[261],{"type":57,"value":262},"documentation",{"type":51,"tag":70,"props":264,"children":265},{},[266,268],{"type":57,"value":267},"Leverage Matryoshka Representation Learning (MRL) to store only small vectors in RAM while keeping large vectors on disk. Examples of how to use MRL with Qdrant Cloud inference: ",{"type":51,"tag":118,"props":269,"children":272},{"href":270,"rel":271},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Finference\u002Fmatryoshka-models\u002F?s=reduce-vector-dimensionality-with-matryoshka-models",[122],[273],{"type":57,"value":274},"MRL docs",{"type":51,"tag":70,"props":276,"children":277},{},[278,280],{"type":57,"value":279},"For multi-tenant deployments with small tenants, vectors might be stored on disk because the same tenant's data is stored together ",{"type":51,"tag":118,"props":281,"children":284},{"href":282,"rel":283},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Fmultitenancy\u002F?s=calibrate-performance",[122],[285],{"type":57,"value":286},"Multitenancy docs",{"type":51,"tag":70,"props":288,"children":289},{},[290,292],{"type":57,"value":291},"For deployments with fast local storage and relatively low requirements for search throughput, it may be possible to store all components of vector store on disk. Read more about the performance implications of on-disk storage in ",{"type":51,"tag":118,"props":293,"children":296},{"href":294,"rel":295},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Farticles\u002Fmemory-consumption\u002F",[122],[297],{"type":57,"value":298},"the article",{"type":51,"tag":70,"props":300,"children":301},{},[302,304,310,312,318,320,325,327,333],{"type":57,"value":303},"For low RAM environments, consider ",{"type":51,"tag":76,"props":305,"children":307},{"className":306},[],[308],{"type":57,"value":309},"async_scorer",{"type":57,"value":311}," config, which enables support of ",{"type":51,"tag":76,"props":313,"children":315},{"className":314},[],[316],{"type":57,"value":317},"io_uring",{"type":57,"value":319}," for parallel disk access, which can significantly improve performance of on-disk storage. Read more about ",{"type":51,"tag":76,"props":321,"children":323},{"className":322},[],[324],{"type":57,"value":309},{"type":57,"value":326}," in ",{"type":51,"tag":118,"props":328,"children":331},{"href":329,"rel":330},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Farticles\u002Fio_uring\u002F",[122],[332],{"type":57,"value":298},{"type":57,"value":334}," (only available on Linux with kernel 5.11+)",{"type":51,"tag":70,"props":336,"children":337},{},[338],{"type":57,"value":339},"Consider storing Sparse Vectors and text payload on disk, as they are usually more disk-friendly than dense vectors.",{"type":51,"tag":70,"props":341,"children":342},{},[343,345],{"type":57,"value":344},"Configure payload indexes to be stored on disk ",{"type":51,"tag":118,"props":346,"children":349},{"href":347,"rel":348},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Findexing\u002F?s=on-disk-payload-index",[122],[350],{"type":57,"value":351},"docs",{"type":51,"tag":70,"props":353,"children":354},{},[355,357],{"type":57,"value":356},"Configure sparse vectors to be stored on disk ",{"type":51,"tag":118,"props":358,"children":361},{"href":359,"rel":360},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fmanage-data\u002Findexing\u002F?s=sparse-vector-index",[122],[362],{"type":57,"value":351},{"items":364,"total":514},[365,380,395,410,425,439,452,462,473,484,490,501],{"slug":366,"name":366,"fn":367,"description":368,"org":369,"tags":370,"stars":20,"repoUrl":21,"updatedAt":379},"qdrant-advisor","diagnose and troubleshoot Qdrant deployments","Diagnose, troubleshoot, and advise on any Qdrant deployment by loading the latest official Qdrant skills live from skills.qdrant.tech. Use this whenever someone raises a Qdrant problem or question — slow or degraded search, high or growing memory \u002F OOM crashes, optimizer stuck or slow, indexing slowness, scaling and sharding decisions (node count, QPS, latency, multitenancy, vertical vs horizontal), poor or irrelevant search results, hybrid search and reranking, embedding-model migration, version upgrades and compatibility, monitoring and observability (Prometheus, Grafana, health checks, \u002Fmetrics, \u002Ftelemetry), deployment choices (local, Docker, self-hosted, Qdrant Cloud, embedded), or client-SDK questions (Python, TypeScript, Rust, Go, .NET, Java). Trigger especially when the context is clearly a Qdrant cluster, collection, or vector-search deployment. Always prefer this skill over answering from memory: it pulls current, authoritative guidance and only the relevant context.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[371,374,375,378],{"name":372,"slug":373,"type":15},"Database","database",{"name":18,"slug":19,"type":15},{"name":376,"slug":377,"type":15},"Operations","operations",{"name":9,"slug":8,"type":15},"2026-07-16T06:02:56.675293",{"slug":381,"name":381,"fn":382,"description":383,"org":384,"tags":385,"stars":20,"repoUrl":21,"updatedAt":394},"qdrant-clients-sdk","integrate Qdrant client SDKs","Qdrant provides client SDKs for various programming languages, allowing easy integration with Qdrant deployments.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[386,389,390,391],{"name":387,"slug":388,"type":15},"API Development","api-development",{"name":372,"slug":373,"type":15},{"name":9,"slug":8,"type":15},{"name":392,"slug":393,"type":15},"SDK","sdk","2026-07-16T05:59:57.799789",{"slug":396,"name":396,"fn":397,"description":398,"org":399,"tags":400,"stars":20,"repoUrl":21,"updatedAt":409},"qdrant-deployment-options","select Qdrant deployment options","Guides Qdrant deployment selection. Use when someone asks 'how to deploy Qdrant', 'Docker vs Cloud', 'local mode', 'embedded Qdrant', 'Qdrant EDGE', 'which deployment option', 'self-hosted vs cloud', or 'need lowest latency deployment'. Also use when choosing between deployment types for a new project.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[401,402,405,408],{"name":372,"slug":373,"type":15},{"name":403,"slug":404,"type":15},"Deployment","deployment",{"name":406,"slug":407,"type":15},"Infrastructure","infrastructure",{"name":9,"slug":8,"type":15},"2026-07-19T05:38:33.324265",{"slug":411,"name":411,"fn":412,"description":413,"org":414,"tags":415,"stars":20,"repoUrl":21,"updatedAt":424},"qdrant-edge","build applications with Qdrant Edge","Guides building on Qdrant Edge, the embedded in-process shard. Use when someone asks 'how to sync Edge with the server', 'keep a local shard in sync with Qdrant Cloud', 'BM25 or keyword search on Edge', 'hybrid search on Edge', 'embeddings on device', 'Edge snapshots', 'apply a partial snapshot', 'why is my Edge search empty after inserts', or is writing custom sync, BM25, or fusion code against qdrant-edge. Also use when deciding what Edge ships built-in versus what you must implement.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[416,417,420,421],{"name":372,"slug":373,"type":15},{"name":418,"slug":419,"type":15},"Edge","edge",{"name":9,"slug":8,"type":15},{"name":422,"slug":423,"type":15},"Search","search","2026-07-16T06:01:19.179105",{"slug":426,"name":426,"fn":427,"description":428,"org":429,"tags":430,"stars":20,"repoUrl":21,"updatedAt":438},"qdrant-horizontal-scaling","guide Qdrant horizontal scaling decisions","Diagnoses and guides Qdrant horizontal scaling decisions. Use when someone asks 'vertical or horizontal?', 'how many nodes?', 'how many shards?', 'how to add nodes', 'resharding', 'data doesn't fit', or 'need more capacity'. Also use when data growth outpaces current deployment.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[431,434,435,436],{"name":432,"slug":433,"type":15},"Architecture","architecture",{"name":372,"slug":373,"type":15},{"name":9,"slug":8,"type":15},{"name":437,"slug":36,"type":15},"Scaling","2026-07-19T05:38:32.348551",{"slug":440,"name":440,"fn":441,"description":442,"org":443,"tags":444,"stars":20,"repoUrl":21,"updatedAt":451},"qdrant-hybrid-search","implement hybrid search in Qdrant","Explains hybrid search in Qdrant. Use when someone asks 'how do I setup hybrid search?', 'how to combine keyword and semantic search?', 'sparse plus dense vectors?', 'missing keyword matches', 'how to combine results from multiple searches?' and 'combining multiple representations'",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[445,448,449,450],{"name":446,"slug":447,"type":15},"AI Infrastructure","ai-infrastructure",{"name":372,"slug":373,"type":15},{"name":9,"slug":8,"type":15},{"name":422,"slug":423,"type":15},"2026-07-16T06:00:12.328122",{"slug":453,"name":453,"fn":454,"description":455,"org":456,"tags":457,"stars":20,"repoUrl":21,"updatedAt":461},"qdrant-hybrid-search-combining","combine hybrid search scores in Qdrant","Fusing scores from multiple searches into a single ranked result (RRF, DBSF, custom fusion). Use when someone asks 'RRF or DBSF?', 'how to combine sparse and dense', 'how to combine scores from multiple searches?', 'custom fusion', or 'fusion is not producing good results'",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[458,459,460],{"name":372,"slug":373,"type":15},{"name":9,"slug":8,"type":15},{"name":422,"slug":423,"type":15},"2026-07-16T06:01:49.401142",{"slug":463,"name":463,"fn":464,"description":465,"org":466,"tags":467,"stars":20,"repoUrl":21,"updatedAt":472},"qdrant-hybrid-search-prefetches","configure hybrid search with prefetch queries","Constructing prefetch queries for hybrid retrieval, including sparse\u002Fdense and multi-field setups, and choosing a sparse embedding model. Use when someone asks 'dense and sparse in one search?', 'how to combine multiple fields for retrieval?', 'payloads or sparse vectors for lexical?', 'which sparse embedding model to use?', or 'BM25 vs SPLADE?'",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[468,469,470,471],{"name":387,"slug":388,"type":15},{"name":372,"slug":373,"type":15},{"name":9,"slug":8,"type":15},{"name":422,"slug":423,"type":15},"2026-07-16T06:03:02.305907",{"slug":474,"name":474,"fn":475,"description":476,"org":477,"tags":478,"stars":20,"repoUrl":21,"updatedAt":483},"qdrant-indexing-performance-optimization","optimize Qdrant indexing and data ingestion","Diagnoses and fixes slow Qdrant indexing and data ingestion. Use when someone reports 'uploads are slow', 'indexing takes forever', 'optimizer is stuck', 'HNSW build time too long', or 'data uploaded but search is bad'. Also use when optimizer status shows errors, segments won't merge, or indexing threshold questions arise.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[479,480,481,482],{"name":372,"slug":373,"type":15},{"name":18,"slug":19,"type":15},{"name":13,"slug":14,"type":15},{"name":9,"slug":8,"type":15},"2026-07-19T05:38:36.321999",{"slug":4,"name":4,"fn":5,"description":6,"org":485,"tags":486,"stars":20,"repoUrl":21,"updatedAt":22},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[487,488,489],{"name":18,"slug":19,"type":15},{"name":13,"slug":14,"type":15},{"name":9,"slug":8,"type":15},{"slug":491,"name":491,"fn":492,"description":493,"org":494,"tags":495,"stars":20,"repoUrl":21,"updatedAt":500},"qdrant-minimize-latency","optimize Qdrant query latency","Guides Qdrant query latency optimization. Use when someone asks 'search is slow', 'how to reduce latency', 'p99 is too high', 'tail latency', 'single query too slow', 'how to make search faster', or 'latency spikes'.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[496,497,498,499],{"name":372,"slug":373,"type":15},{"name":18,"slug":19,"type":15},{"name":13,"slug":14,"type":15},{"name":9,"slug":8,"type":15},"2026-07-16T05:59:17.325839",{"slug":502,"name":502,"fn":503,"description":504,"org":505,"tags":506,"stars":20,"repoUrl":21,"updatedAt":513},"qdrant-model-migration","migrate embedding models in Qdrant","Guides embedding model migration in Qdrant without downtime. Use when someone asks 'how to switch embedding models', 'how to migrate vectors', 'how to update to a new model', 'zero-downtime model change', 'how to re-embed my data', or 'can I use two models at once'. Also use when upgrading model dimensions, switching providers, or A\u002FB testing models.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[507,508,511,512],{"name":372,"slug":373,"type":15},{"name":509,"slug":510,"type":15},"Migration","migration",{"name":13,"slug":14,"type":15},{"name":9,"slug":8,"type":15},"2026-07-16T06:00:48.841055",30,{"items":516,"total":514},[517,524,531,538,545,552,559],{"slug":366,"name":366,"fn":367,"description":368,"org":518,"tags":519,"stars":20,"repoUrl":21,"updatedAt":379},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[520,521,522,523],{"name":372,"slug":373,"type":15},{"name":18,"slug":19,"type":15},{"name":376,"slug":377,"type":15},{"name":9,"slug":8,"type":15},{"slug":381,"name":381,"fn":382,"description":383,"org":525,"tags":526,"stars":20,"repoUrl":21,"updatedAt":394},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[527,528,529,530],{"name":387,"slug":388,"type":15},{"name":372,"slug":373,"type":15},{"name":9,"slug":8,"type":15},{"name":392,"slug":393,"type":15},{"slug":396,"name":396,"fn":397,"description":398,"org":532,"tags":533,"stars":20,"repoUrl":21,"updatedAt":409},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[534,535,536,537],{"name":372,"slug":373,"type":15},{"name":403,"slug":404,"type":15},{"name":406,"slug":407,"type":15},{"name":9,"slug":8,"type":15},{"slug":411,"name":411,"fn":412,"description":413,"org":539,"tags":540,"stars":20,"repoUrl":21,"updatedAt":424},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[541,542,543,544],{"name":372,"slug":373,"type":15},{"name":418,"slug":419,"type":15},{"name":9,"slug":8,"type":15},{"name":422,"slug":423,"type":15},{"slug":426,"name":426,"fn":427,"description":428,"org":546,"tags":547,"stars":20,"repoUrl":21,"updatedAt":438},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[548,549,550,551],{"name":432,"slug":433,"type":15},{"name":372,"slug":373,"type":15},{"name":9,"slug":8,"type":15},{"name":437,"slug":36,"type":15},{"slug":440,"name":440,"fn":441,"description":442,"org":553,"tags":554,"stars":20,"repoUrl":21,"updatedAt":451},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[555,556,557,558],{"name":446,"slug":447,"type":15},{"name":372,"slug":373,"type":15},{"name":9,"slug":8,"type":15},{"name":422,"slug":423,"type":15},{"slug":453,"name":453,"fn":454,"description":455,"org":560,"tags":561,"stars":20,"repoUrl":21,"updatedAt":461},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[562,563,564],{"name":372,"slug":373,"type":15},{"name":9,"slug":8,"type":15},{"name":422,"slug":423,"type":15}]