[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-qdrant-qdrant-hybrid-search-combining":3,"mdc-mzbwmw-key":46,"related-org-qdrant-qdrant-hybrid-search-combining":405,"related-repo-qdrant-qdrant-hybrid-search-combining":556},{"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":41,"sourceUrl":44,"mdContent":45},"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},"qdrant","Qdrant","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fqdrant.png",[12,14,17],{"name":9,"slug":8,"type":13},"tag",{"name":15,"slug":16,"type":13},"Database","database",{"name":18,"slug":19,"type":13},"Search","search",197,"https:\u002F\u002Fgithub.com\u002Fqdrant\u002Fskills","2026-07-16T06:01:49.401142",null,23,[26,27,28,29,30,31,32,33,34,35,8,36,37,38,39,40],"agent-skills","ai-agents","claude-code","codex","cursor","embeddings","hybrid-search","monitoring","multitenancy","performance","quantization","scaling","search-quality","vector-database","vector-search",{"repoUrl":21,"stars":20,"forks":24,"topics":42,"description":43},[26,27,28,29,30,31,32,33,34,35,8,36,37,38,39,40],"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-search-quality\u002Fsearch-strategies\u002Fhybrid-search\u002Fcombining-searches","---\nname: qdrant-hybrid-search-combining\ndescription: \"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'\"\n---\n\n# Combining Prefetch Results\n\nThe outer query fuses ranked candidate lists from all parallel prefetches into one ranked list of results. Fusion methods differ in whether they use rank, score or directly vector representations of candidates (their similarity to the outer query) and whether final score incorporates payload metadata. All methods support flat (one fusion step) and nested (multi-stage) prefetch structures.\n\n## Scores Are Not Comparable Across Prefetches & You Want Some Easy Baseline\n\nUse when: searches produce scores on different scales, like BM25 and cosine on dense embeddings.\n\n### RRF\n- **[RRF](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fsearch\u002Fhybrid-queries\u002F?s=reciprocal-rank-fusion-rrf)** (Reciprocal Rank Fusion) — rank-based, ignores scores magnitude, a decent default to start with.\n- Tune `k` to [control rank sensitivity in RRF fusion](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fsearch\u002Fhybrid-queries\u002F?s=setting-rrf-constant-k).\n- Add per-prefetch **weights** when one search should dominate, using [Weighted RRF](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fsearch\u002Fhybrid-queries\u002F?s=weighted-rrf). Weights should be customized per collection and retrievers' score distributions!\n\n### DBSF\n- **[DBSF](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fsearch\u002Fhybrid-queries\u002F?s=distribution-based-score-fusion-dbsf)** (Distribution-Based Score Fusion) — normalizes score distributions per prefetch before fusing them, for that, instead of min-max, uses mean +- 3 deviations on prefetched list of scores. Avoid relying on resulting absolute scores, as scores in DBSF are normalized per prefetch (aka per a retrieved list of search results), and might be uncomparable across queries.\n\n## Need Custom Fusion\n\nUse when: recency, popularity or other payload values should affect the merged ranking alongside candidate scores or you need a custom fusion.\n\n**[With formula query](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fsearch\u002Fsearch-relevance\u002F?s=score-boosting)**, access `score` of each prefetch and, if desired, payload field values.\n\nIf you want to implement custom fusion on `score` of each prefetch:\n- Use decay or any other available expressions for normalizing score distributions before fusing them. \n- Parameters of these expressions should be based on the collection & retriever score distributions (for example, adjusting these parameters on a subsample of real queries). \n- Formula query is unable to provide ranks for custom fusions \n\nWhen using `FormulaQuery` over multiple prefetches (e.g. per-representation weighting):\n- `$score[i]` indexes prefetches in declaration order. Reordering the `prefetch=` list silently shifts which weight applies to which retriever.\n- Provide `defaults` for every `$score[i]` so the formula still evaluates for candidates that surfaced from only a subset of prefetches.\n- Start with RRF when scores are on incomparable scales (e.g. BM25 + cosine). Reach for `FormulaQuery` only when explicit per-representation weighting or payload-driven boosts are required, and normalize each `$score[i]` (decay or min-max on a sampled distribution) before combining linearly.\n\n## Need Good Ranking of Fused Candidates and Ready To Spend More Resources\n\nUse when: you want to use similarity between query and candidates' vector representations as the prefetches combiner and simultaneously ranker. \nMore resource heavy than score\u002Frank based fusions, but might be necessary due to use case requirements or need in a high top-K precision of results (when parallel prefetches have overall a good recall of retrieved candidates).\n\nYou can use any type of vector as an outer query over the prefetches, to perform the fusion on the server-side in one QueryAPI request: sparse, dense, multivector. For that, same type of vector representations for documents need to be stored as named vectors per point.\n\nInstead of using client-side fusion through cross-encoders, a popular option is **Late interaction models-based fusion**, through reranking on multivectors (e.g. ColBERT for text, ColPali and ColQwen for images).\n- Most precise but highest compute\u002Fresource usage.\n- Configure multivectors used for fusion through reranking with HNSW disabled like in [Hybrid Search with Reranking tutorial](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Ftutorials-basics\u002Freranking-hybrid-search\u002F).\n\n## What NOT to Do\n\n- Use linear weighted fusion on incomparable score ranges. [Why not](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Farticles\u002Fhybrid-search\u002F?s=why-not-a-linear-combination).\n- Use \"vibe\" defined weights in weighted RRF. Weights should be fine-tuned per dataset and retrieval pipelines.\n- Pick any fusion type without comparative experiments.\n- Use late interaction multivectors for fusion without evaluating cheaper analogues, for example, MUVERA. More in [multi-vector Qdrant search course](https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fcourse\u002Fmulti-vector-search\u002F)\n",{"data":47,"body":48},{"name":4,"description":6},{"type":49,"children":50},"root",[51,60,66,73,78,85,151,157,174,180,185,208,220,238,251,312,318,323,328,340,361,367],{"type":52,"tag":53,"props":54,"children":56},"element","h1",{"id":55},"combining-prefetch-results",[57],{"type":58,"value":59},"text","Combining Prefetch Results",{"type":52,"tag":61,"props":62,"children":63},"p",{},[64],{"type":58,"value":65},"The outer query fuses ranked candidate lists from all parallel prefetches into one ranked list of results. Fusion methods differ in whether they use rank, score or directly vector representations of candidates (their similarity to the outer query) and whether final score incorporates payload metadata. All methods support flat (one fusion step) and nested (multi-stage) prefetch structures.",{"type":52,"tag":67,"props":68,"children":70},"h2",{"id":69},"scores-are-not-comparable-across-prefetches-you-want-some-easy-baseline",[71],{"type":58,"value":72},"Scores Are Not Comparable Across Prefetches & You Want Some Easy Baseline",{"type":52,"tag":61,"props":74,"children":75},{},[76],{"type":58,"value":77},"Use when: searches produce scores on different scales, like BM25 and cosine on dense embeddings.",{"type":52,"tag":79,"props":80,"children":82},"h3",{"id":81},"rrf",[83],{"type":58,"value":84},"RRF",{"type":52,"tag":86,"props":87,"children":88},"ul",{},[89,107,130],{"type":52,"tag":90,"props":91,"children":92},"li",{},[93,105],{"type":52,"tag":94,"props":95,"children":96},"strong",{},[97],{"type":52,"tag":98,"props":99,"children":103},"a",{"href":100,"rel":101},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fsearch\u002Fhybrid-queries\u002F?s=reciprocal-rank-fusion-rrf",[102],"nofollow",[104],{"type":58,"value":84},{"type":58,"value":106}," (Reciprocal Rank Fusion) — rank-based, ignores scores magnitude, a decent default to start with.",{"type":52,"tag":90,"props":108,"children":109},{},[110,112,119,121,128],{"type":58,"value":111},"Tune ",{"type":52,"tag":113,"props":114,"children":116},"code",{"className":115},[],[117],{"type":58,"value":118},"k",{"type":58,"value":120}," to ",{"type":52,"tag":98,"props":122,"children":125},{"href":123,"rel":124},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fsearch\u002Fhybrid-queries\u002F?s=setting-rrf-constant-k",[102],[126],{"type":58,"value":127},"control rank sensitivity in RRF fusion",{"type":58,"value":129},".",{"type":52,"tag":90,"props":131,"children":132},{},[133,135,140,142,149],{"type":58,"value":134},"Add per-prefetch ",{"type":52,"tag":94,"props":136,"children":137},{},[138],{"type":58,"value":139},"weights",{"type":58,"value":141}," when one search should dominate, using ",{"type":52,"tag":98,"props":143,"children":146},{"href":144,"rel":145},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fsearch\u002Fhybrid-queries\u002F?s=weighted-rrf",[102],[147],{"type":58,"value":148},"Weighted RRF",{"type":58,"value":150},". Weights should be customized per collection and retrievers' score distributions!",{"type":52,"tag":79,"props":152,"children":154},{"id":153},"dbsf",[155],{"type":58,"value":156},"DBSF",{"type":52,"tag":86,"props":158,"children":159},{},[160],{"type":52,"tag":90,"props":161,"children":162},{},[163,172],{"type":52,"tag":94,"props":164,"children":165},{},[166],{"type":52,"tag":98,"props":167,"children":170},{"href":168,"rel":169},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fsearch\u002Fhybrid-queries\u002F?s=distribution-based-score-fusion-dbsf",[102],[171],{"type":58,"value":156},{"type":58,"value":173}," (Distribution-Based Score Fusion) — normalizes score distributions per prefetch before fusing them, for that, instead of min-max, uses mean +- 3 deviations on prefetched list of scores. Avoid relying on resulting absolute scores, as scores in DBSF are normalized per prefetch (aka per a retrieved list of search results), and might be uncomparable across queries.",{"type":52,"tag":67,"props":175,"children":177},{"id":176},"need-custom-fusion",[178],{"type":58,"value":179},"Need Custom Fusion",{"type":52,"tag":61,"props":181,"children":182},{},[183],{"type":58,"value":184},"Use when: recency, popularity or other payload values should affect the merged ranking alongside candidate scores or you need a custom fusion.",{"type":52,"tag":61,"props":186,"children":187},{},[188,198,200,206],{"type":52,"tag":94,"props":189,"children":190},{},[191],{"type":52,"tag":98,"props":192,"children":195},{"href":193,"rel":194},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Fsearch\u002Fsearch-relevance\u002F?s=score-boosting",[102],[196],{"type":58,"value":197},"With formula query",{"type":58,"value":199},", access ",{"type":52,"tag":113,"props":201,"children":203},{"className":202},[],[204],{"type":58,"value":205},"score",{"type":58,"value":207}," of each prefetch and, if desired, payload field values.",{"type":52,"tag":61,"props":209,"children":210},{},[211,213,218],{"type":58,"value":212},"If you want to implement custom fusion on ",{"type":52,"tag":113,"props":214,"children":216},{"className":215},[],[217],{"type":58,"value":205},{"type":58,"value":219}," of each prefetch:",{"type":52,"tag":86,"props":221,"children":222},{},[223,228,233],{"type":52,"tag":90,"props":224,"children":225},{},[226],{"type":58,"value":227},"Use decay or any other available expressions for normalizing score distributions before fusing them.",{"type":52,"tag":90,"props":229,"children":230},{},[231],{"type":58,"value":232},"Parameters of these expressions should be based on the collection & retriever score distributions (for example, adjusting these parameters on a subsample of real queries).",{"type":52,"tag":90,"props":234,"children":235},{},[236],{"type":58,"value":237},"Formula query is unable to provide ranks for custom fusions",{"type":52,"tag":61,"props":239,"children":240},{},[241,243,249],{"type":58,"value":242},"When using ",{"type":52,"tag":113,"props":244,"children":246},{"className":245},[],[247],{"type":58,"value":248},"FormulaQuery",{"type":58,"value":250}," over multiple prefetches (e.g. per-representation weighting):",{"type":52,"tag":86,"props":252,"children":253},{},[254,273,293],{"type":52,"tag":90,"props":255,"children":256},{},[257,263,265,271],{"type":52,"tag":113,"props":258,"children":260},{"className":259},[],[261],{"type":58,"value":262},"$score[i]",{"type":58,"value":264}," indexes prefetches in declaration order. Reordering the ",{"type":52,"tag":113,"props":266,"children":268},{"className":267},[],[269],{"type":58,"value":270},"prefetch=",{"type":58,"value":272}," list silently shifts which weight applies to which retriever.",{"type":52,"tag":90,"props":274,"children":275},{},[276,278,284,286,291],{"type":58,"value":277},"Provide ",{"type":52,"tag":113,"props":279,"children":281},{"className":280},[],[282],{"type":58,"value":283},"defaults",{"type":58,"value":285}," for every ",{"type":52,"tag":113,"props":287,"children":289},{"className":288},[],[290],{"type":58,"value":262},{"type":58,"value":292}," so the formula still evaluates for candidates that surfaced from only a subset of prefetches.",{"type":52,"tag":90,"props":294,"children":295},{},[296,298,303,305,310],{"type":58,"value":297},"Start with RRF when scores are on incomparable scales (e.g. BM25 + cosine). Reach for ",{"type":52,"tag":113,"props":299,"children":301},{"className":300},[],[302],{"type":58,"value":248},{"type":58,"value":304}," only when explicit per-representation weighting or payload-driven boosts are required, and normalize each ",{"type":52,"tag":113,"props":306,"children":308},{"className":307},[],[309],{"type":58,"value":262},{"type":58,"value":311}," (decay or min-max on a sampled distribution) before combining linearly.",{"type":52,"tag":67,"props":313,"children":315},{"id":314},"need-good-ranking-of-fused-candidates-and-ready-to-spend-more-resources",[316],{"type":58,"value":317},"Need Good Ranking of Fused Candidates and Ready To Spend More Resources",{"type":52,"tag":61,"props":319,"children":320},{},[321],{"type":58,"value":322},"Use when: you want to use similarity between query and candidates' vector representations as the prefetches combiner and simultaneously ranker.\nMore resource heavy than score\u002Frank based fusions, but might be necessary due to use case requirements or need in a high top-K precision of results (when parallel prefetches have overall a good recall of retrieved candidates).",{"type":52,"tag":61,"props":324,"children":325},{},[326],{"type":58,"value":327},"You can use any type of vector as an outer query over the prefetches, to perform the fusion on the server-side in one QueryAPI request: sparse, dense, multivector. For that, same type of vector representations for documents need to be stored as named vectors per point.",{"type":52,"tag":61,"props":329,"children":330},{},[331,333,338],{"type":58,"value":332},"Instead of using client-side fusion through cross-encoders, a popular option is ",{"type":52,"tag":94,"props":334,"children":335},{},[336],{"type":58,"value":337},"Late interaction models-based fusion",{"type":58,"value":339},", through reranking on multivectors (e.g. ColBERT for text, ColPali and ColQwen for images).",{"type":52,"tag":86,"props":341,"children":342},{},[343,348],{"type":52,"tag":90,"props":344,"children":345},{},[346],{"type":58,"value":347},"Most precise but highest compute\u002Fresource usage.",{"type":52,"tag":90,"props":349,"children":350},{},[351,353,360],{"type":58,"value":352},"Configure multivectors used for fusion through reranking with HNSW disabled like in ",{"type":52,"tag":98,"props":354,"children":357},{"href":355,"rel":356},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fdocumentation\u002Ftutorials-basics\u002Freranking-hybrid-search\u002F",[102],[358],{"type":58,"value":359},"Hybrid Search with Reranking tutorial",{"type":58,"value":129},{"type":52,"tag":67,"props":362,"children":364},{"id":363},"what-not-to-do",[365],{"type":58,"value":366},"What NOT to Do",{"type":52,"tag":86,"props":368,"children":369},{},[370,383,388,393],{"type":52,"tag":90,"props":371,"children":372},{},[373,375,382],{"type":58,"value":374},"Use linear weighted fusion on incomparable score ranges. ",{"type":52,"tag":98,"props":376,"children":379},{"href":377,"rel":378},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Farticles\u002Fhybrid-search\u002F?s=why-not-a-linear-combination",[102],[380],{"type":58,"value":381},"Why not",{"type":58,"value":129},{"type":52,"tag":90,"props":384,"children":385},{},[386],{"type":58,"value":387},"Use \"vibe\" defined weights in weighted RRF. Weights should be fine-tuned per dataset and retrieval pipelines.",{"type":52,"tag":90,"props":389,"children":390},{},[391],{"type":58,"value":392},"Pick any fusion type without comparative experiments.",{"type":52,"tag":90,"props":394,"children":395},{},[396,398],{"type":58,"value":397},"Use late interaction multivectors for fusion without evaluating cheaper analogues, for example, MUVERA. More in ",{"type":52,"tag":98,"props":399,"children":402},{"href":400,"rel":401},"https:\u002F\u002Fskills.qdrant.tech\u002Fmd\u002Fcourse\u002Fmulti-vector-search\u002F",[102],[403],{"type":58,"value":404},"multi-vector Qdrant search course",{"items":406,"total":555},[407,422,437,452,465,479,492,498,509,521,531,542],{"slug":408,"name":408,"fn":409,"description":410,"org":411,"tags":412,"stars":20,"repoUrl":21,"updatedAt":421},"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},[413,414,417,420],{"name":15,"slug":16,"type":13},{"name":415,"slug":416,"type":13},"Debugging","debugging",{"name":418,"slug":419,"type":13},"Operations","operations",{"name":9,"slug":8,"type":13},"2026-07-16T06:02:56.675293",{"slug":423,"name":423,"fn":424,"description":425,"org":426,"tags":427,"stars":20,"repoUrl":21,"updatedAt":436},"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},[428,431,432,433],{"name":429,"slug":430,"type":13},"API Development","api-development",{"name":15,"slug":16,"type":13},{"name":9,"slug":8,"type":13},{"name":434,"slug":435,"type":13},"SDK","sdk","2026-07-16T05:59:57.799789",{"slug":438,"name":438,"fn":439,"description":440,"org":441,"tags":442,"stars":20,"repoUrl":21,"updatedAt":451},"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},[443,444,447,450],{"name":15,"slug":16,"type":13},{"name":445,"slug":446,"type":13},"Deployment","deployment",{"name":448,"slug":449,"type":13},"Infrastructure","infrastructure",{"name":9,"slug":8,"type":13},"2026-07-19T05:38:33.324265",{"slug":453,"name":453,"fn":454,"description":455,"org":456,"tags":457,"stars":20,"repoUrl":21,"updatedAt":464},"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},[458,459,462,463],{"name":15,"slug":16,"type":13},{"name":460,"slug":461,"type":13},"Edge","edge",{"name":9,"slug":8,"type":13},{"name":18,"slug":19,"type":13},"2026-07-16T06:01:19.179105",{"slug":466,"name":466,"fn":467,"description":468,"org":469,"tags":470,"stars":20,"repoUrl":21,"updatedAt":478},"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},[471,474,475,476],{"name":472,"slug":473,"type":13},"Architecture","architecture",{"name":15,"slug":16,"type":13},{"name":9,"slug":8,"type":13},{"name":477,"slug":37,"type":13},"Scaling","2026-07-19T05:38:32.348551",{"slug":480,"name":480,"fn":481,"description":482,"org":483,"tags":484,"stars":20,"repoUrl":21,"updatedAt":491},"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},[485,488,489,490],{"name":486,"slug":487,"type":13},"AI Infrastructure","ai-infrastructure",{"name":15,"slug":16,"type":13},{"name":9,"slug":8,"type":13},{"name":18,"slug":19,"type":13},"2026-07-16T06:00:12.328122",{"slug":4,"name":4,"fn":5,"description":6,"org":493,"tags":494,"stars":20,"repoUrl":21,"updatedAt":22},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[495,496,497],{"name":15,"slug":16,"type":13},{"name":9,"slug":8,"type":13},{"name":18,"slug":19,"type":13},{"slug":499,"name":499,"fn":500,"description":501,"org":502,"tags":503,"stars":20,"repoUrl":21,"updatedAt":508},"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},[504,505,506,507],{"name":429,"slug":430,"type":13},{"name":15,"slug":16,"type":13},{"name":9,"slug":8,"type":13},{"name":18,"slug":19,"type":13},"2026-07-16T06:03:02.305907",{"slug":510,"name":510,"fn":511,"description":512,"org":513,"tags":514,"stars":20,"repoUrl":21,"updatedAt":520},"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},[515,516,517,519],{"name":15,"slug":16,"type":13},{"name":415,"slug":416,"type":13},{"name":518,"slug":35,"type":13},"Performance",{"name":9,"slug":8,"type":13},"2026-07-19T05:38:36.321999",{"slug":522,"name":522,"fn":523,"description":524,"org":525,"tags":526,"stars":20,"repoUrl":21,"updatedAt":530},"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},[527,528,529],{"name":415,"slug":416,"type":13},{"name":518,"slug":35,"type":13},{"name":9,"slug":8,"type":13},"2026-07-16T06:00:15.357243",{"slug":532,"name":532,"fn":533,"description":534,"org":535,"tags":536,"stars":20,"repoUrl":21,"updatedAt":541},"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},[537,538,539,540],{"name":15,"slug":16,"type":13},{"name":415,"slug":416,"type":13},{"name":518,"slug":35,"type":13},{"name":9,"slug":8,"type":13},"2026-07-16T05:59:17.325839",{"slug":543,"name":543,"fn":544,"description":545,"org":546,"tags":547,"stars":20,"repoUrl":21,"updatedAt":554},"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},[548,549,552,553],{"name":15,"slug":16,"type":13},{"name":550,"slug":551,"type":13},"Migration","migration",{"name":518,"slug":35,"type":13},{"name":9,"slug":8,"type":13},"2026-07-16T06:00:48.841055",30,{"items":557,"total":555},[558,565,572,579,586,593,600],{"slug":408,"name":408,"fn":409,"description":410,"org":559,"tags":560,"stars":20,"repoUrl":21,"updatedAt":421},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[561,562,563,564],{"name":15,"slug":16,"type":13},{"name":415,"slug":416,"type":13},{"name":418,"slug":419,"type":13},{"name":9,"slug":8,"type":13},{"slug":423,"name":423,"fn":424,"description":425,"org":566,"tags":567,"stars":20,"repoUrl":21,"updatedAt":436},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[568,569,570,571],{"name":429,"slug":430,"type":13},{"name":15,"slug":16,"type":13},{"name":9,"slug":8,"type":13},{"name":434,"slug":435,"type":13},{"slug":438,"name":438,"fn":439,"description":440,"org":573,"tags":574,"stars":20,"repoUrl":21,"updatedAt":451},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[575,576,577,578],{"name":15,"slug":16,"type":13},{"name":445,"slug":446,"type":13},{"name":448,"slug":449,"type":13},{"name":9,"slug":8,"type":13},{"slug":453,"name":453,"fn":454,"description":455,"org":580,"tags":581,"stars":20,"repoUrl":21,"updatedAt":464},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[582,583,584,585],{"name":15,"slug":16,"type":13},{"name":460,"slug":461,"type":13},{"name":9,"slug":8,"type":13},{"name":18,"slug":19,"type":13},{"slug":466,"name":466,"fn":467,"description":468,"org":587,"tags":588,"stars":20,"repoUrl":21,"updatedAt":478},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[589,590,591,592],{"name":472,"slug":473,"type":13},{"name":15,"slug":16,"type":13},{"name":9,"slug":8,"type":13},{"name":477,"slug":37,"type":13},{"slug":480,"name":480,"fn":481,"description":482,"org":594,"tags":595,"stars":20,"repoUrl":21,"updatedAt":491},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[596,597,598,599],{"name":486,"slug":487,"type":13},{"name":15,"slug":16,"type":13},{"name":9,"slug":8,"type":13},{"name":18,"slug":19,"type":13},{"slug":4,"name":4,"fn":5,"description":6,"org":601,"tags":602,"stars":20,"repoUrl":21,"updatedAt":22},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[603,604,605],{"name":15,"slug":16,"type":13},{"name":9,"slug":8,"type":13},{"name":18,"slug":19,"type":13}]