{
  "listings": [
    {
      "slug": "weaviate",
      "name": "Weaviate API + MCP",
      "company": "Weaviate B.V.",
      "grade": "C",
      "fit": "yes",
      "why": "weaviate-collections-get-config shows property names, weaviate-tenants-list lists tenants, and weaviate-query-hybrid takes text and filters; a viewer-role key or the read-only switch keeps it read-only.",
      "source": "https://docs.weaviate.io/weaviate/configuration/mcp-server.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/weaviate"
    },
    {
      "slug": "chroma",
      "name": "Chroma API + MCP",
      "company": "Chroma Inc.",
      "grade": "D",
      "fit": "caveat",
      "why": "chroma-mcp lists collections with paging, returns collection info and runs text queries with filters, but it has no read-only mode, so read-only credentials depend on the client used.",
      "source": "https://raw.githubusercontent.com/chroma-core/chroma-mcp/main/README.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/chroma"
    },
    {
      "slug": "cloudflare-ai-search",
      "name": "Cloudflare AI Search",
      "company": "Cloudflare, Inc.",
      "grade": "BB",
      "fit": "caveat",
      "why": "The per-instance MCP endpoint exposes one search tool, so the agent cannot list instances or fields, and the endpoint is public once enabled unless something like Cloudflare Access sits in front.",
      "source": "https://developers.cloudflare.com/ai-search/api/search/mcp/index.md",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/cloudflare-ai-search"
    },
    {
      "slug": "elasticsearch",
      "name": "Elasticsearch",
      "company": "Elasticsearch B.V.",
      "grade": "C",
      "fit": "caveat",
      "why": "The Agent Builder MCP server exposes built-in and custom tools, limited by the authorising user's permissions, but it needs Kibana Agent Builder (Elastic Cloud or 9.2 and later) and some built-in tools are unavailable over MCP.",
      "source": "https://www.elastic.co/docs/solutions/search/agent-builder/mcp-server",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/elasticsearch"
    },
    {
      "slug": "meilisearch",
      "name": "Meilisearch",
      "company": "Meili SAS",
      "grade": "A",
      "fit": "caveat",
      "why": "The built-in MCP route has four read-only tools, including searchInIndexes with filters and hybrid, but it is experimental and off by default, and describeIndex returns five sample documents rather than a field list.",
      "source": "https://www.meilisearch.com/docs/getting_started/integrations/mcp.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/meilisearch"
    },
    {
      "slug": "milvus-zilliz",
      "name": "Milvus and Zilliz Cloud API + MCP",
      "company": "Zilliz Inc.",
      "grade": "B",
      "fit": "caveat",
      "why": "The MCP server lists collections, returns schema with milvus_get_collection_info and runs text search with filter_expr, but it has no read-only mode and also has create, load, release and delete tools.",
      "source": "https://raw.githubusercontent.com/zilliztech/mcp-server-milvus/main/README.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/milvus-zilliz"
    },
    {
      "slug": "opensearch",
      "name": "OpenSearch",
      "company": "OpenSearch Software Foundation, a project of The Linux Foundation",
      "grade": "BB",
      "fit": "caveat",
      "why": "Core MCP tools list indices, return mappings and run DSL searches, but the GenericOpenSearchApiTool can call any endpoint, #43 shows a wrong field returning an empty result the model trusted, and #203 filters ignored in multi mode.",
      "source": "https://raw.githubusercontent.com/opensearch-project/opensearch-mcp-server-py/main/README.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/opensearch"
    },
    {
      "slug": "pinecone",
      "name": "Pinecone API + MCP",
      "company": "Pinecone Systems Inc.",
      "grade": "A",
      "fit": "caveat",
      "why": "MCP tools list and describe indexes and search with text and metadata filters, but only on legacy integrated-embedding indexes, and no read-only mode was found.",
      "source": "https://docs.pinecone.io/guides/agents/mcp-server.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/pinecone"
    },
    {
      "slug": "qdrant",
      "name": "Qdrant API + MCP",
      "company": "Qdrant Solutions GmbH",
      "grade": "BB",
      "fit": "caveat",
      "why": "The MCP server takes text queries and has QDRANT_READ_ONLY, but its two tools cannot list collections or fields, and filters need extra configuration.",
      "source": "https://raw.githubusercontent.com/qdrant/mcp-server-qdrant/master/README.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/qdrant"
    },
    {
      "slug": "typesense",
      "name": "Typesense API + MCP",
      "company": "Typesense, Inc.",
      "grade": "BB",
      "fit": "caveat",
      "why": "The Typesense Cloud MCP server runs keyword and hybrid searches with permissions chosen at connect, but first connect is a browser consent page, so a headless agent cannot start alone.",
      "source": "https://typesense.org/docs/guide/typesense-cloud/mcp-server.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/typesense"
    },
    {
      "slug": "upstash-vector",
      "name": "Upstash Vector API + MCP",
      "company": "Upstash, Inc.",
      "grade": "B",
      "fit": "caveat",
      "why": "The hosted MCP server lists indexes and namespaces, runs vector_query with server-side embedding, and a read-only OAuth grant or key removes write tools, but no tool lists metadata field names.",
      "source": "https://upstash.com/docs/agent-resources/mcp.md",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/upstash-vector"
    },
    {
      "slug": "epsilla",
      "name": "Epsilla Vector Database",
      "company": "Epsilla, Inc.",
      "grade": "E",
      "fit": "no",
      "why": "No MCP server found in the docs index.",
      "source": "https://epsilla-inc.gitbook.io/llms.txt",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/epsilla"
    },
    {
      "slug": "lancedb",
      "name": "LanceDB",
      "company": "LanceDB Systems, Inc.",
      "grade": "B",
      "fit": "no",
      "why": "No official MCP server or hosted endpoint for an agent to call without an Enterprise deployment.",
      "source": "https://docs.lancedb.com/llms.txt",
      "confidence": "high",
      "url": "https://www.anchorterminal.com/tools/lancedb"
    },
    {
      "slug": "vespa",
      "name": "Vespa",
      "company": "Vespa.ai AS",
      "grade": "BB",
      "fit": "no",
      "why": "No official public MCP server; an MCP endpoint must be built into the user's own application.",
      "source": "https://docs.vespa.ai/llms.txt",
      "confidence": "medium",
      "url": "https://www.anchorterminal.com/tools/vespa"
    }
  ],
  "marks": "yes: the docs show it does the job as the success test says. caveat: it does the job with a limit an agent should know. no: it doesn't do this job. unchecked: we couldn't read it.",
  "meta": {
    "attribution": "Anchor Terminal (https://www.anchorterminal.com)",
    "attributionRequired": true,
    "credit": "CC BY 4.0, credit Anchor Terminal and link the page or API address you took it from (https://www.anchorterminal.com/terms/#data)",
    "docs": "https://www.anchorterminal.com/docs/",
    "generatedAt": "2026-11-16",
    "license": "CC-BY-4.0",
    "method": "https://www.anchorterminal.com/benchmark/",
    "methodology": "0.8",
    "openapi": "https://www.anchorterminal.com/openapi.json",
    "preview": false,
    "run": "2026-10-01",
    "runLabel": "October 2026 research run",
    "terms": "https://www.anchorterminal.com/terms/#data"
  },
  "useCase": {
    "accounts": 5,
    "after": [
      "vector-search.rag-over-docs"
    ],
    "area": "developer",
    "asks": [
      "search the vector store for X",
      "what collections are there",
      "find records where status is open about Y"
    ],
    "before": [],
    "blockers": [
      "data tools need a browser sign-in so a headless agent gets cluster tools only",
      "MCP server lags the engine (no hybrid, no filters without extra flags)"
    ],
    "capabilities": [
      "db.vector",
      "db.filters",
      "db.hybrid"
    ],
    "category": "vector-search",
    "categoryName": "Retrieval \u0026 vector search",
    "checked": "2026-11-15",
    "companies": {
      "fit": 1,
      "fitWithLimit": 10,
      "tagged": 18
    },
    "competition": "one",
    "fits": [
      "weaviate"
    ],
    "fitsWithLimit": [
      "chroma",
      "cloudflare-ai-search",
      "elasticsearch",
      "meilisearch",
      "milvus-zilliz",
      "opensearch",
      "pinecone",
      "qdrant",
      "typesense",
      "upstash-vector"
    ],
    "id": "vector-search.agent-query-tool",
    "listings": {
      "caveat": 10,
      "no": 3,
      "notRead": 0,
      "tagged": 18,
      "unchecked": 0,
      "yes": 1
    },
    "name": "Let an agent query an index through MCP without writing code",
    "status": "draft",
    "success": "The agent lists collections and their fields, sends a text query with a filter on a real field, and gets results, without building a vector or guessing a field name, using read-only credentials. Measured: the first useful result arrives within three tool calls.",
    "url": "https://www.anchorterminal.com/api/v1/use-cases/vector-search.agent-query-tool.json",
    "weight": 6,
    "weightBasis": "5 of 37 accounts on MCP servers (Milvus, OpenSearch, Meilisearch, Chroma, Pinecone)"
  }
}
