[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6QFQWFA1HDBWJSC8MGDZJ8J":3,"skill-same:01M6QFQWFA1HDBWJSC8MGDZJ8J":87},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":39,"linkOnly":40,"markdownContent":41,"markdownBody":42,"frontMatter":43,"refName":50,"commitSha":51,"skillDirectory":52,"forksCount":14,"repositoryHtmlUrl":53,"repositoryHomepageUrl":54,"ownerHtmlUrl":55,"ownerBlogUrl":56,"ownerDescription":57,"updatedAt":58,"install":59,"resources":72},"01M6QFQWFA1HDBWJSC8MGDZJ8J","qdrant-vector-search",null,"High-performance vector similarity search engine for RAG and semantic search. Use when building production RAG systems requiring fast nearest neighbor search, hybrid search with filtering, or scalable vector storage with Rust-powered performance.","Tashima-Tarsh\u002FDisha6.6","Tashima-Tarsh","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F269906498?v=4","01M6PBH98X03XERH0HFT4PFB1P","NITISH KUMAR","01M6PBH9DA0M6KM3Z12P848F0K",0,"MIT","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fkal-qdrant-vector-search\u002FSKILL.md",[],[19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38],"search","vector","Vector Search","Qdrant","qdrant vector search","filtering","hybrid","for","high","Hybrid Search","client","high-performance","systems","use","distributed","engine","knowledge","embeddings","semantic search","rag","2026-11-10 10:12:30",false,"---\nname: qdrant-vector-search\ndescription: High-performance vector similarity search engine for RAG and semantic search. Use when building production RAG systems requiring fast nearest neighbor search, hybrid search with filtering, or scalable vector storage with Rust-powered performance.\nlicense: MIT\nmetadata:\n  version: 1.0.0\n  category: knowledge-and-rag\n  maintainer: Kalaris Labs\n  tags: RAG, Vector Search, Qdrant, Semantic Search, Embeddings, Similarity Search, HNSW, Production, Distributed\n  dependencies: qdrant-client>=1.12.0\n---\n\n# Qdrant - Vector Similarity Search Engine\n\nHigh-performance vector database written in Rust for production RAG and semantic search.\n\n## When to use Qdrant\n\n**Use Qdrant when:**\n- Building production RAG systems requiring low latency\n- Need hybrid search (vectors + metadata filtering)\n- Require horizontal scaling with sharding\u002Freplication\n- Want on-premise deployment with full data control\n- Need multi-vector storage per record (dense + sparse)\n- Building real-time recommendation systems\n\n**Key features:**\n- **Rust-powered**: Memory-safe, high performance\n- **Rich filtering**: Filter by any payload field during search\n- **Multiple vectors**: Dense, sparse, multi-dense per point\n- **Quantization**: Scalar, product, binary for memory efficiency\n- **Distributed**: Raft consensus, sharding, replication\n- **REST + gRPC**: Both APIs with full feature parity\n\n**Use alternatives instead:**\n- **Chroma**: Simpler setup, embedded use cases\n- **FAISS**: Maximum raw speed, research\u002Fbatch processing\n- **Pinecone**: Fully managed, zero ops preferred\n- **Weaviate**: GraphQL preference, built-in vectorizers\n\n## Quick start\n\n### Installation\n\n```bash\n# Python client\npip install qdrant-client\n\n# Docker (recommended for development)\ndocker run -p 6333:6333 -p 6334:6334 qdrant\u002Fqdrant\n\n# Docker with persistent storage\ndocker run -p 6333:6333 -p 6334:6334 \\\n    -v $(pwd)\u002Fqdrant_storage:\u002Fqdrant\u002Fstorage \\\n    qdrant\u002Fqdrant\n```\n\n### Basic usage\n\n```python\nfrom qdrant_client import QdrantClient\nfrom qdrant_client.models import Distance, VectorParams, PointStruct\n\n# Connect to Qdrant\nclient = QdrantClient(host=\"localhost\", port=6333)\n\n# Create collection\nclient.create_collection(\n    collection_name=\"documents\",\n    vectors_config=VectorParams(size=384, distance=Distance.COSINE)\n)\n\n# Insert vectors with payload\nclient.upsert(\n    collection_name=\"documents\",\n    points=[\n        PointStruct(\n            id=1,\n            vector=[0.1, 0.2, ...],  # 384-dim vector\n            payload={\"title\": \"Doc 1\", \"category\": \"tech\"}\n        ),\n        PointStruct(\n            id=2,\n            vector=[0.3, 0.4, ...],\n            payload={\"title\": \"Doc 2\", \"category\": \"science\"}\n        )\n    ]\n)\n\n# Search with filtering\nresults = client.search(\n    collection_name=\"documents\",\n    query_vector=[0.15, 0.25, ...],\n    query_filter={\n        \"must\": [{\"key\": \"category\", \"match\": {\"value\": \"tech\"}}]\n    },\n    limit=10\n)\n\nfor point in results:\n    print(f\"ID: {point.id}, Score: {point.score}, Payload: {point.payload}\")\n```\n\n## Core concepts\n\n### Points - Basic data unit\n\n```python\nfrom qdrant_client.models import PointStruct\n\n# Point = ID + Vector(s) + Payload\npoint = PointStruct(\n    id=123,                              # Integer or UUID string\n    vector=[0.1, 0.2, 0.3, ...],        # Dense vector\n    payload={                            # Arbitrary JSON metadata\n        \"title\": \"Document title\",\n        \"category\": \"tech\",\n        \"timestamp\": 1699900000,\n        \"tags\": [\"python\", \"ml\"]\n    }\n)\n\n# Batch upsert (recommended)\nclient.upsert(\n    collection_name=\"documents\",\n    points=[point1, point2, point3],\n    wait=True  # Wait for indexing\n)\n```\n\n### Collections - Vector containers\n\n```python\nfrom qdrant_client.models import VectorParams, Distance, HnswConfigDiff\n\n# Create with HNSW configuration\nclient.create_collection(\n    collection_name=\"documents\",\n    vectors_config=VectorParams(\n        size=384,                        # Vector dimensions\n        distance=Distance.COSINE         # COSINE, EUCLID, DOT, MANHATTAN\n    ),\n    hnsw_config=HnswConfigDiff(\n        m=16,                            # Connections per node (default 16)\n        ef_construct=100,                # Build-time accuracy (default 100)\n        full_scan_threshold=10000        # Switch to brute force below this\n    ),\n    on_disk_payload=True                 # Store payload on disk\n)\n\n# Collection info\ninfo = client.get_collection(\"documents\")\nprint(f\"Points: {info.points_count}, Vectors: {info.vectors_count}\")\n```\n\n### Distance metrics\n\n| Metric | Use Case | Range |\n|--------|----------|-------|\n| `COSINE` | Text embeddings, normalized vectors | 0 to 2 |\n| `EUCLID` | Spatial data, image features | 0 to ∞ |\n| `DOT` | Recommendations, unnormalized | -∞ to ∞ |\n| `MANHATTAN` | Sparse features, discrete data | 0 to ∞ |\n\n## Search operations\n\n### Basic search\n\n```python\n# Simple nearest neighbor search\nresults = client.search(\n    collection_name=\"documents\",\n    query_vector=[0.1, 0.2, ...],\n    limit=10,\n    with_payload=True,\n    with_vectors=False  # Don't return vectors (faster)\n)\n```\n\n### Filtered search\n\n```python\nfrom qdrant_client.models import Filter, FieldCondition, MatchValue, Range\n\n# Complex filtering\nresults = client.search(\n    collection_name=\"documents\",\n    query_vector=query_embedding,\n    query_filter=Filter(\n        must=[\n            FieldCondition(key=\"category\", match=MatchValue(value=\"tech\")),\n            FieldCondition(key=\"timestamp\", range=Range(gte=1699000000))\n        ],\n        must_not=[\n            FieldCondition(key=\"status\", match=MatchValue(value=\"archived\"))\n        ]\n    ),\n    limit=10\n)\n\n# Shorthand filter syntax\nresults = client.search(\n    collection_name=\"documents\",\n    query_vector=query_embedding,\n    query_filter={\n        \"must\": [\n            {\"key\": \"category\", \"match\": {\"value\": \"tech\"}},\n            {\"key\": \"price\", \"range\": {\"gte\": 10, \"lte\": 100}}\n        ]\n    },\n    limit=10\n)\n```\n\n### Batch search\n\n```python\nfrom qdrant_client.models import SearchRequest\n\n# Multiple queries in one request\nresults = client.search_batch(\n    collection_name=\"documents\",\n    requests=[\n        SearchRequest(vector=[0.1, ...], limit=5),\n        SearchRequest(vector=[0.2, ...], limit=5, filter={\"must\": [...]}),\n        SearchRequest(vector=[0.3, ...], limit=10)\n    ]\n)\n```\n\n## RAG integration\n\nDetails, code examples and parameter tables: [references\u002Frag-integration.md](references\u002Frag-integration.md). Read it when this step applies.\n\n## Multi-vector support\n\n### Named vectors (different embedding models)\n\n```python\nfrom qdrant_client.models import VectorParams, Distance\n\n# Collection with multiple vector types\nclient.create_collection(\n    collection_name=\"hybrid_search\",\n    vectors_config={\n        \"dense\": VectorParams(size=384, distance=Distance.COSINE),\n        \"sparse\": VectorParams(size=30000, distance=Distance.DOT)\n    }\n)\n\n# Insert with named vectors\nclient.upsert(\n    collection_name=\"hybrid_search\",\n    points=[\n        PointStruct(\n            id=1,\n            vector={\n                \"dense\": dense_embedding,\n                \"sparse\": sparse_embedding\n            },\n            payload={\"text\": \"document text\"}\n        )\n    ]\n)\n\n# Search specific vector\nresults = client.search(\n    collection_name=\"hybrid_search\",\n    query_vector=(\"dense\", query_dense),  # Specify which vector\n    limit=10\n)\n```\n\n### Sparse vectors (BM25, SPLADE)\n\n```python\nfrom qdrant_client.models import SparseVectorParams, SparseIndexParams, SparseVector\n\n# Collection with sparse vectors\nclient.create_collection(\n    collection_name=\"sparse_search\",\n    vectors_config={},\n    sparse_vectors_config={\"text\": SparseVectorParams(index=SparseIndexParams(on_disk=False))}\n)\n\n# Insert sparse vector\nclient.upsert(\n    collection_name=\"sparse_search\",\n    points=[PointStruct(id=1, vector={\"text\": SparseVector(indices=[1, 5, 100], values=[0.5, 0.8, 0.2])}, payload={\"text\": \"document\"})]\n)\n```\n\n## Quantization (memory optimization)\n\n```python\nfrom qdrant_client.models import ScalarQuantization, ScalarQuantizationConfig, ScalarType\n\n# Scalar quantization (4x memory reduction)\nclient.create_collection(\n    collection_name=\"quantized\",\n    vectors_config=VectorParams(size=384, distance=Distance.COSINE),\n    quantization_config=ScalarQuantization(\n        scalar=ScalarQuantizationConfig(\n            type=ScalarType.INT8,\n            quantile=0.99,        # Clip outliers\n            always_ram=True      # Keep quantized in RAM\n        )\n    )\n)\n\n# Search with rescoring\nresults = client.search(\n    collection_name=\"quantized\",\n    query_vector=query,\n    search_params={\"quantization\": {\"rescore\": True}},  # Rescore top results\n    limit=10\n)\n```\n\n## Payload indexing\n\n```python\nfrom qdrant_client.models import PayloadSchemaType\n\n# Create payload index for faster filtering\nclient.create_payload_index(\n    collection_name=\"documents\",\n    field_name=\"category\",\n    field_schema=PayloadSchemaType.KEYWORD\n)\n\nclient.create_payload_index(\n    collection_name=\"documents\",\n    field_name=\"timestamp\",\n    field_schema=PayloadSchemaType.INTEGER\n)\n\n# Index types: KEYWORD, INTEGER, FLOAT, GEO, TEXT (full-text), BOOL\n```\n\n## Production deployment\n\n### Qdrant Cloud\n\n```python\nfrom qdrant_client import QdrantClient\n\n# Connect to Qdrant Cloud\nclient = QdrantClient(\n    url=\"https:\u002F\u002Fyour-cluster.cloud.qdrant.io\",\n    api_key=\"your-api-key\"\n)\n```\n\n### Performance tuning\n\n```python\n# Optimize for search speed (higher recall)\nclient.update_collection(\n    collection_name=\"documents\",\n    hnsw_config=HnswConfigDiff(ef_construct=200, m=32)\n)\n\n# Optimize for indexing speed (bulk loads)\nclient.update_collection(\n    collection_name=\"documents\",\n    optimizer_config={\"indexing_threshold\": 20000}\n)\n```\n\n## Best practices\n\n1. **Batch operations** - Use batch upsert\u002Fsearch for efficiency\n2. **Payload indexing** - Index fields used in filters\n3. **Quantization** - Enable for large collections (>1M vectors)\n4. **Sharding** - Use for collections >10M vectors\n5. **On-disk storage** - Enable `on_disk_payload` for large payloads\n6. **Connection pooling** - Reuse client instances\n\n## Common issues\n\n**Slow search with filters:**\n```python\n# Create payload index for filtered fields\nclient.create_payload_index(\n    collection_name=\"docs\",\n    field_name=\"category\",\n    field_schema=PayloadSchemaType.KEYWORD\n)\n```\n\n**Out of memory:**\n```python\n# Enable quantization and on-disk storage\nclient.create_collection(\n    collection_name=\"large_collection\",\n    vectors_config=VectorParams(size=384, distance=Distance.COSINE),\n    quantization_config=ScalarQuantization(...),\n    on_disk_payload=True\n)\n```\n\n**Connection issues:**\n```python\n# Use timeout and retry\nclient = QdrantClient(\n    host=\"localhost\",\n    port=6333,\n    timeout=30,\n    prefer_grpc=True  # gRPC for better performance\n)\n```\n\n## References\n\n- **[Advanced Usage](references\u002Fadvanced-usage.md)** - Distributed mode, hybrid search, recommendations\n- **[Troubleshooting](references\u002Ftroubleshooting.md)** - Common issues, debugging, performance tuning\n\n## Resources\n\n- **GitHub**: https:\u002F\u002Fgithub.com\u002Fqdrant\u002Fqdrant\n- **Docs**: https:\u002F\u002Fqdrant.tech\u002Fdocumentation\u002F\n- **Python Client**: https:\u002F\u002Fgithub.com\u002Fqdrant\u002Fqdrant-client\n- **Cloud**: https:\u002F\u002Fcloud.qdrant.io\n- **Version**: 1.12.0+\n- **License**: Apache 2.0\n\n## Agent operating procedure\n\n1. **Check the environment.** Confirm the corpus, embedding model and vector store versions, and where the index will live.\n2. **Pin down the inputs.** Confirm formats, identifiers and parameters from the data or the user. Ask rather than guess any value that changes the result.\n3. **Run a small version first.** Index a small subset and test retrieval on known question-passage pairs.\n4. **Execute the full task** using the instructions and references above.\n5. **Validate the result.** Measure recall@k on a labeled set; answers cite retrieved passages only.\n6. **Report.** State what was run (versions, commands, parameters), what was checked, and what is still uncertain.\n\n| If this happens | Do this |\n|---|---|\n| Retrieval returns irrelevant chunks | Revisit chunking, add BM25\u002Fhybrid search, or re-rank before changing the generator. |\n| A function, flag or endpoint in these instructions is missing in the installed version | Check the installed version's own documentation (`help()`, `--help`, official docs), adapt, and tell the user. Never invent an API. |\n| A required input, identifier or parameter is ambiguous | Ask the user, or state the assumption explicitly before running. |\n\n**Integrity rules**\n\n- Never fabricate results, parameters, identifiers, citations or statistics. If something cannot be run or verified, say so plainly.\n- Answers must come from retrieved text; say when the corpus does not contain the answer.\n- Treat version-specific details here as possibly outdated: confirm them against the official documentation for the installed version.\n- Ask before actions that cost money, consume shared GPUs or cloud quota, touch personal or patient data, or cannot be undone.\n\n## Related skills\n\n- `pinecone`: Guides use of Pinecone, a managed serverless vector database, through its Python client and the LangChain and LlamaIndex integrations.\n- `faiss`: Facebook's library for efficient similarity search and clustering of dense vectors.\n- `chroma`: Open-source embedding database for AI applications.\n","# Qdrant - Vector Similarity Search Engine\n\nHigh-performance vector database written in Rust for production RAG and semantic search.\n\n## When to use Qdrant\n\n**Use Qdrant when:**\n- Building production RAG systems requiring low latency\n- Need hybrid search (vectors + metadata filtering)\n- Require horizontal scaling with sharding\u002Freplication\n- Want on-premise deployment with full data control\n- Need multi-vector storage per record (dense + sparse)\n- Building real-time recommendation systems\n\n**Key features:**\n- **Rust-powered**: Memory-safe, high performance\n- **Rich filtering**: Filter by any payload field during search\n- **Multiple vectors**: Dense, sparse, multi-dense per point\n- **Quantization**: Scalar, product, binary for memory efficiency\n- **Distributed**: Raft consensus, sharding, replication\n- **REST + gRPC**: Both APIs with full feature parity\n\n**Use alternatives instead:**\n- **Chroma**: Simpler setup, embedded use cases\n- **FAISS**: Maximum raw speed, research\u002Fbatch processing\n- **Pinecone**: Fully managed, zero ops preferred\n- **Weaviate**: GraphQL preference, built-in vectorizers\n\n## Quick start\n\n### Installation\n\n```bash\n# Python client\npip install qdrant-client\n\n# Docker (recommended for development)\ndocker run -p 6333:6333 -p 6334:6334 qdrant\u002Fqdrant\n\n# Docker with persistent storage\ndocker run -p 6333:6333 -p 6334:6334 \\\n    -v $(pwd)\u002Fqdrant_storage:\u002Fqdrant\u002Fstorage \\\n    qdrant\u002Fqdrant\n```\n\n### Basic usage\n\n```python\nfrom qdrant_client import QdrantClient\nfrom qdrant_client.models import Distance, VectorParams, PointStruct\n\n# Connect to Qdrant\nclient = QdrantClient(host=\"localhost\", port=6333)\n\n# Create collection\nclient.create_collection(\n    collection_name=\"documents\",\n    vectors_config=VectorParams(size=384, distance=Distance.COSINE)\n)\n\n# Insert vectors with payload\nclient.upsert(\n    collection_name=\"documents\",\n    points=[\n        PointStruct(\n            id=1,\n            vector=[0.1, 0.2, ...],  # 384-dim vector\n            payload={\"title\": \"Doc 1\", \"category\": \"tech\"}\n        ),\n        PointStruct(\n            id=2,\n            vector=[0.3, 0.4, ...],\n            payload={\"title\": \"Doc 2\", \"category\": \"science\"}\n        )\n    ]\n)\n\n# Search with filtering\nresults = client.search(\n    collection_name=\"documents\",\n    query_vector=[0.15, 0.25, ...],\n    query_filter={\n        \"must\": [{\"key\": \"category\", \"match\": {\"value\": \"tech\"}}]\n    },\n    limit=10\n)\n\nfor point in results:\n    print(f\"ID: {point.id}, Score: {point.score}, Payload: {point.payload}\")\n```\n\n## Core concepts\n\n### Points - Basic data unit\n\n```python\nfrom qdrant_client.models import PointStruct\n\n# Point = ID + Vector(s) + Payload\npoint = PointStruct(\n    id=123,                              # Integer or UUID string\n    vector=[0.1, 0.2, 0.3, ...],        # Dense vector\n    payload={                            # Arbitrary JSON metadata\n        \"title\": \"Document title\",\n        \"category\": \"tech\",\n        \"timestamp\": 1699900000,\n        \"tags\": [\"python\", \"ml\"]\n    }\n)\n\n# Batch upsert (recommended)\nclient.upsert(\n    collection_name=\"documents\",\n    points=[point1, point2, point3],\n    wait=True  # Wait for indexing\n)\n```\n\n### Collections - Vector containers\n\n```python\nfrom qdrant_client.models import VectorParams, Distance, HnswConfigDiff\n\n# Create with HNSW configuration\nclient.create_collection(\n    collection_name=\"documents\",\n    vectors_config=VectorParams(\n        size=384,                        # Vector dimensions\n        distance=Distance.COSINE         # COSINE, EUCLID, DOT, MANHATTAN\n    ),\n    hnsw_config=HnswConfigDiff(\n        m=16,                            # Connections per node (default 16)\n        ef_construct=100,                # Build-time accuracy (default 100)\n        full_scan_threshold=10000        # Switch to brute force below this\n    ),\n    on_disk_payload=True                 # Store payload on disk\n)\n\n# Collection info\ninfo = client.get_collection(\"documents\")\nprint(f\"Points: {info.points_count}, Vectors: {info.vectors_count}\")\n```\n\n### Distance metrics\n\n| Metric | Use Case | Range |\n|--------|----------|-------|\n| `COSINE` | Text embeddings, normalized vectors | 0 to 2 |\n| `EUCLID` | Spatial data, image features | 0 to ∞ |\n| `DOT` | Recommendations, unnormalized | -∞ to ∞ |\n| `MANHATTAN` | Sparse features, discrete data | 0 to ∞ |\n\n## Search operations\n\n### Basic search\n\n```python\n# Simple nearest neighbor search\nresults = client.search(\n    collection_name=\"documents\",\n    query_vector=[0.1, 0.2, ...],\n    limit=10,\n    with_payload=True,\n    with_vectors=False  # Don't return vectors (faster)\n)\n```\n\n### Filtered search\n\n```python\nfrom qdrant_client.models import Filter, FieldCondition, MatchValue, Range\n\n# Complex filtering\nresults = client.search(\n    collection_name=\"documents\",\n    query_vector=query_embedding,\n    query_filter=Filter(\n        must=[\n            FieldCondition(key=\"category\", match=MatchValue(value=\"tech\")),\n            FieldCondition(key=\"timestamp\", range=Range(gte=1699000000))\n        ],\n        must_not=[\n            FieldCondition(key=\"status\", match=MatchValue(value=\"archived\"))\n        ]\n    ),\n    limit=10\n)\n\n# Shorthand filter syntax\nresults = client.search(\n    collection_name=\"documents\",\n    query_vector=query_embedding,\n    query_filter={\n        \"must\": [\n            {\"key\": \"category\", \"match\": {\"value\": \"tech\"}},\n            {\"key\": \"price\", \"range\": {\"gte\": 10, \"lte\": 100}}\n        ]\n    },\n    limit=10\n)\n```\n\n### Batch search\n\n```python\nfrom qdrant_client.models import SearchRequest\n\n# Multiple queries in one request\nresults = client.search_batch(\n    collection_name=\"documents\",\n    requests=[\n        SearchRequest(vector=[0.1, ...], limit=5),\n        SearchRequest(vector=[0.2, ...], limit=5, filter={\"must\": [...]}),\n        SearchRequest(vector=[0.3, ...], limit=10)\n    ]\n)\n```\n\n## RAG integration\n\nDetails, code examples and parameter tables: [references\u002Frag-integration.md](references\u002Frag-integration.md). Read it when this step applies.\n\n## Multi-vector support\n\n### Named vectors (different embedding models)\n\n```python\nfrom qdrant_client.models import VectorParams, Distance\n\n# Collection with multiple vector types\nclient.create_collection(\n    collection_name=\"hybrid_search\",\n    vectors_config={\n        \"dense\": VectorParams(size=384, distance=Distance.COSINE),\n        \"sparse\": VectorParams(size=30000, distance=Distance.DOT)\n    }\n)\n\n# Insert with named vectors\nclient.upsert(\n    collection_name=\"hybrid_search\",\n    points=[\n        PointStruct(\n            id=1,\n            vector={\n                \"dense\": dense_embedding,\n                \"sparse\": sparse_embedding\n            },\n            payload={\"text\": \"document text\"}\n        )\n    ]\n)\n\n# Search specific vector\nresults = client.search(\n    collection_name=\"hybrid_search\",\n    query_vector=(\"dense\", query_dense),  # Specify which vector\n    limit=10\n)\n```\n\n### Sparse vectors (BM25, SPLADE)\n\n```python\nfrom qdrant_client.models import SparseVectorParams, SparseIndexParams, SparseVector\n\n# Collection with sparse vectors\nclient.create_collection(\n    collection_name=\"sparse_search\",\n    vectors_config={},\n    sparse_vectors_config={\"text\": SparseVectorParams(index=SparseIndexParams(on_disk=False))}\n)\n\n# Insert sparse vector\nclient.upsert(\n    collection_name=\"sparse_search\",\n    points=[PointStruct(id=1, vector={\"text\": SparseVector(indices=[1, 5, 100], values=[0.5, 0.8, 0.2])}, payload={\"text\": \"document\"})]\n)\n```\n\n## Quantization (memory optimization)\n\n```python\nfrom qdrant_client.models import ScalarQuantization, ScalarQuantizationConfig, ScalarType\n\n# Scalar quantization (4x memory reduction)\nclient.create_collection(\n    collection_name=\"quantized\",\n    vectors_config=VectorParams(size=384, distance=Distance.COSINE),\n    quantization_config=ScalarQuantization(\n        scalar=ScalarQuantizationConfig(\n            type=ScalarType.INT8,\n            quantile=0.99,        # Clip outliers\n            always_ram=True      # Keep quantized in RAM\n        )\n    )\n)\n\n# Search with rescoring\nresults = client.search(\n    collection_name=\"quantized\",\n    query_vector=query,\n    search_params={\"quantization\": {\"rescore\": True}},  # Rescore top results\n    limit=10\n)\n```\n\n## Payload indexing\n\n```python\nfrom qdrant_client.models import PayloadSchemaType\n\n# Create payload index for faster filtering\nclient.create_payload_index(\n    collection_name=\"documents\",\n    field_name=\"category\",\n    field_schema=PayloadSchemaType.KEYWORD\n)\n\nclient.create_payload_index(\n    collection_name=\"documents\",\n    field_name=\"timestamp\",\n    field_schema=PayloadSchemaType.INTEGER\n)\n\n# Index types: KEYWORD, INTEGER, FLOAT, GEO, TEXT (full-text), BOOL\n```\n\n## Production deployment\n\n### Qdrant Cloud\n\n```python\nfrom qdrant_client import QdrantClient\n\n# Connect to Qdrant Cloud\nclient = QdrantClient(\n    url=\"https:\u002F\u002Fyour-cluster.cloud.qdrant.io\",\n    api_key=\"your-api-key\"\n)\n```\n\n### Performance tuning\n\n```python\n# Optimize for search speed (higher recall)\nclient.update_collection(\n    collection_name=\"documents\",\n    hnsw_config=HnswConfigDiff(ef_construct=200, m=32)\n)\n\n# Optimize for indexing speed (bulk loads)\nclient.update_collection(\n    collection_name=\"documents\",\n    optimizer_config={\"indexing_threshold\": 20000}\n)\n```\n\n## Best practices\n\n1. **Batch operations** - Use batch upsert\u002Fsearch for efficiency\n2. **Payload indexing** - Index fields used in filters\n3. **Quantization** - Enable for large collections (>1M vectors)\n4. **Sharding** - Use for collections >10M vectors\n5. **On-disk storage** - Enable `on_disk_payload` for large payloads\n6. **Connection pooling** - Reuse client instances\n\n## Common issues\n\n**Slow search with filters:**\n```python\n# Create payload index for filtered fields\nclient.create_payload_index(\n    collection_name=\"docs\",\n    field_name=\"category\",\n    field_schema=PayloadSchemaType.KEYWORD\n)\n```\n\n**Out of memory:**\n```python\n# Enable quantization and on-disk storage\nclient.create_collection(\n    collection_name=\"large_collection\",\n    vectors_config=VectorParams(size=384, distance=Distance.COSINE),\n    quantization_config=ScalarQuantization(...),\n    on_disk_payload=True\n)\n```\n\n**Connection issues:**\n```python\n# Use timeout and retry\nclient = QdrantClient(\n    host=\"localhost\",\n    port=6333,\n    timeout=30,\n    prefer_grpc=True  # gRPC for better performance\n)\n```\n\n## References\n\n- **[Advanced Usage](references\u002Fadvanced-usage.md)** - Distributed mode, hybrid search, recommendations\n- **[Troubleshooting](references\u002Ftroubleshooting.md)** - Common issues, debugging, performance tuning\n\n## Resources\n\n- **GitHub**: https:\u002F\u002Fgithub.com\u002Fqdrant\u002Fqdrant\n- **Docs**: https:\u002F\u002Fqdrant.tech\u002Fdocumentation\u002F\n- **Python Client**: https:\u002F\u002Fgithub.com\u002Fqdrant\u002Fqdrant-client\n- **Cloud**: https:\u002F\u002Fcloud.qdrant.io\n- **Version**: 1.12.0+\n- **License**: Apache 2.0\n\n## Agent operating procedure\n\n1. **Check the environment.** Confirm the corpus, embedding model and vector store versions, and where the index will live.\n2. **Pin down the inputs.** Confirm formats, identifiers and parameters from the data or the user. Ask rather than guess any value that changes the result.\n3. **Run a small version first.** Index a small subset and test retrieval on known question-passage pairs.\n4. **Execute the full task** using the instructions and references above.\n5. **Validate the result.** Measure recall@k on a labeled set; answers cite retrieved passages only.\n6. **Report.** State what was run (versions, commands, parameters), what was checked, and what is still uncertain.\n\n| If this happens | Do this |\n|---|---|\n| Retrieval returns irrelevant chunks | Revisit chunking, add BM25\u002Fhybrid search, or re-rank before changing the generator. |\n| A function, flag or endpoint in these instructions is missing in the installed version | Check the installed version's own documentation (`help()`, `--help`, official docs), adapt, and tell the user. Never invent an API. |\n| A required input, identifier or parameter is ambiguous | Ask the user, or state the assumption explicitly before running. |\n\n**Integrity rules**\n\n- Never fabricate results, parameters, identifiers, citations or statistics. If something cannot be run or verified, say so plainly.\n- Answers must come from retrieved text; say when the corpus does not contain the answer.\n- Treat version-specific details here as possibly outdated: confirm them against the official documentation for the installed version.\n- Ask before actions that cost money, consume shared GPUs or cloud quota, touch personal or patient data, or cannot be undone.\n\n## Related skills\n\n- `pinecone`: Guides use of Pinecone, a managed serverless vector database, through its Python client and the LangChain and LlamaIndex integrations.\n- `faiss`: Facebook's library for efficient similarity search and clustering of dense vectors.\n- `chroma`: Open-source embedding database for AI applications.\n",{"name":5,"license":15,"metadata":44,"description":7},{"tags":45,"version":46,"category":47,"maintainer":48,"dependencies":49},"RAG, Vector Search, Qdrant, Semantic Search, Embeddings, Similarity Search, HNSW, Production, Distributed","1.0.0","knowledge-and-rag","Kalaris Labs","qdrant-client>=1.12.0","main","0da4485785ee599388900be50db1635489d53056",".agents\u002Fskills\u002Fkal-qdrant-vector-search","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6","https:\u002F\u002Fdisha6.6.thenitishkr.in\u002F","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh","www.thenitishkr","thenitishkr | AI • Cyber-security • Research\r\n","2026-11-08 03:33:25",{"source":8,"skillName":5,"commands":60,"prompt":70,"zipAvailable":71},[61,64,67],{"tool":62,"command":63},"npx","npx skills add Tashima-Tarsh\u002FDisha6.6 --skill qdrant-vector-search",{"tool":65,"command":66},"bunx","bunx skills add Tashima-Tarsh\u002FDisha6.6 --skill qdrant-vector-search",{"tool":68,"command":69},"pnpm","pnpm dlx skills add Tashima-Tarsh\u002FDisha6.6 --skill qdrant-vector-search","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fkal-qdrant-vector-search\u002FSKILL.md",true,[73,79,83],{"path":74,"resourceType":75,"resourceTypeDesc":76,"sizeBytes":77,"sourceHtmlUrl":78},"references\u002Fadvanced-usage.md","REFERENCE","参考文档",14969,"https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fkal-qdrant-vector-search\u002Freferences\u002Fadvanced-usage.md",{"path":80,"resourceType":75,"resourceTypeDesc":76,"sizeBytes":81,"sourceHtmlUrl":82},"references\u002Frag-integration.md",2234,"https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fkal-qdrant-vector-search\u002Freferences\u002Frag-integration.md",{"path":84,"resourceType":75,"resourceTypeDesc":76,"sizeBytes":85,"sourceHtmlUrl":86},"references\u002Ftroubleshooting.md",12682,"https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fkal-qdrant-vector-search\u002Freferences\u002Ftroubleshooting.md",[88,118,141,168,193,217],{"skillNo":89,"name":90,"nameZh":6,"description":91,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":92,"categories":93,"keywords":99,"publishedAt":39},"01M6QRD07D4TGNDK6VJ4EHF8BV","youtube-full","Use when the user needs YouTube transcripts, video search, channel browsing, playlist extraction, or content monitoring. Trigger phrases: 'get the transcript for', 'search YouTube for', 'what are the latest videos on', 'list this playlist', 'monitor this channel', or any request involving a YouTube URL, video ID, or @handle. Do NOT use for downloading video or audio files, YouTube engagement data (likes, comments), or private\u002Fage-restricted videos.","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fyoutube-full\u002FSKILL.md",[94],{"categoryNo":95,"categoryName":96,"taxonomyType":97,"taxonomyTypeDesc":98},"01JBSEEDCAT000000000000016","图像与多媒体","SCENARIO","使用场景",[100,101,102,103,19,104,105,106,107,108,109,110,111,112,113,114,26,115,116,117],"youtube full","youtube","full","transcript","monitoring","data","Video","Audio","list","files","do","get","age","trigger","URL","browsing","request","youtube transcripts",{"skillNo":119,"name":120,"nameZh":6,"description":121,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":122,"categories":123,"keywords":124,"publishedAt":39},"01M6QRCJ3112VYPA97YG7KGMC2","x-twitter-growth","X\u002FTwitter growth engine for building audience, crafting viral content, and analyzing engagement. Use when the user wants to grow on X\u002FTwitter, write tweets or threads, analyze their X profile, research competitors on X, plan a posting strategy, or optimize engagement. Complements social-content (generic multi-platform) with X-specific depth: algorithm mechanics, thread engineering, reply strategy, profile optimization, and competitive intelligence via web search.","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fx-twitter-growth\u002FSKILL.md",[],[125,126,127,128,19,129,130,131,132,133,134,34,135,136,32,137,138,139,140,26],"twitter","x twitter growth","growth","research","content","engineering","plan","web search","competitors","analyze","write","strategy","platform","grow","posting","competitive",{"skillNo":142,"name":143,"nameZh":6,"description":144,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":145,"categories":146,"keywords":150,"publishedAt":39},"01M6QRBN3EHQNJY42RERP7A149","write-a-skill","Create new agent skills with proper structure, progressive disclosure, and bundled resources. Use when user wants to create, write, build, or author a new skill.","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fwrite-a-skill\u002FSKILL.md",[147],{"categoryNo":148,"categoryName":149,"taxonomyType":97,"taxonomyTypeDesc":98},"01JBSEEDCAT000000000000023","Skill与智能体开发",[151,135,152,153,154,155,156,157,32,158,159,50,160,161,162,163,164,165,166,167],"write a skill","github","create","https","productivity","matt pocock","direct","example","build","new skill","structure","new agent","resources","with","tree","progressive disclosure","author",{"skillNo":169,"name":170,"nameZh":6,"description":171,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":172,"categories":173,"keywords":174,"publishedAt":39},"01M6QRB6G8P33PNGAF9Y69EJ9T","wrap-up","Close out a launched Claude Managed Agent — recap every primitive the founder now owns, regenerate the single-file overview page, and suggest the next 1-2 upgrades. Use when the user says \"wrap up\", \"close this out\", \"what do I own now\", \"give me the summary\", \"recap the agent\", or when the orchestrator routes phase=wrap-up. primitives_inventory.py tables everything owned (agent, environment, session, memory, outcome, deployment); overview_page.py regenerates a self-contained .\u002Fmy-agent\u002Fagent-overview.html; upgrade_suggester.py ranks the next moves from recorded deferrals plus standing hardening steps. Companion to run-without-you; the last stop before phase=done.","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fwrap-up\u002FSKILL.md",[],[175,176,177,178,179,180,181,182,183,32,184,110,185,186,187,188,189,190,191,192],"wrap up","up","wrap","gemini","deployment","run","file","hardening","closeout","stop","upgrade","founder","close","companion","you","tables","directions","phase",{"skillNo":194,"name":195,"nameZh":6,"description":196,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":197,"categories":198,"keywords":200,"publishedAt":39},"01M6QRA5AYC1RBFZ39GTWGS3N8","workflow-builder","Design and write deterministic multi-agent workflow scripts (.js files in .claude\u002Fworkflows\u002F) for Claude Code's Workflow tool. Use when a user wants to build, create, author, scaffold, or run a custom Claude Code workflow, orchestrate sub-agents (fan-out, pipeline, loop, judge-panel), or automate a repeatable multi-step task across fresh-context agents.","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fworkflow-builder\u002FSKILL.md",[199],{"categoryNo":148,"categoryName":149,"taxonomyType":97,"taxonomyTypeDesc":98},[201,202,203,204,205,153,180,206,207,208,209,210,135,211,109,212,213,214,215,216],"workflow","workflow builder","builder","design","workflows","code","orchestrate","pipeline","loop","multi-agent","context","across","fresh","scripts","judge","claude code workflow",{"skillNo":218,"name":219,"nameZh":6,"description":220,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":221,"categories":222,"keywords":223,"publishedAt":39},"01M6QR7QMCZESKQA0ERW1E657M","weekly-review","Use when someone wants to run a weekly review, close open loops, audit stalled projects and commitments, get their system back to trusted, restart a lapsed review habit, or says \"\u002Fcs:weekly-review\". Walks David Allen's three-phase loop — GET CLEAR, GET CURRENT, GET CREATIVE — with deterministic scripts that inventory open loops, gate the checklist with named gaps, and score commitment health 0-100.","https:\u002F\u002Fgithub.com\u002FTashima-Tarsh\u002FDisha6.6\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fweekly-review\u002FSKILL.md",[],[224,225,226,227,228,229,230,231,232,209,233,234,235,236,237,238,111,239,240,192],"review","weekly review","weekly","deterministic scripts","reflect","feeds","inventory","open","team","projects","project-management","scanner","conversation","back","project","sprint","cadence"]