mcp-server-weaviate

iteam1/mcp-server-weaviate

3.2

If you are the rightful owner of mcp-server-weaviate and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to dayong@mcphub.com.

The Model Context Protocol (MCP) server for Weaviate is designed to facilitate seamless integration and communication between machine learning models and the Weaviate vector database.

mcp-server-weaviate

📋 Prerequisites

  • uv
  • weaviate
  • LLM provider API key (OpenAI, Anthropic, etc.) or local LLM (e.g. Ollama)

🛠️ Tools

1. test_connection

Test connection to Weaviate server and get server information (version, hostname, available modules).

2. list_collections

List all available collections in Weaviate with record counts and sample properties.

3. near_text_search

Perform semantic/vector search in a collection.

  • Parameters:
    • collection_name (required): Name of the collection to search in
    • query_text (required): Text query to search for semantically similar results
    • limit (optional): Maximum number of results to return (default: 5)

4. keyword_search

Perform keyword search (BM25) in a collection.

  • Parameters:
    • collection_name (required): Name of the collection to search in
    • query_text (required): Keywords to search for using BM25 algorithm
    • limit (optional): Maximum number of results to return (default: 5)

5. hybrid_search

Perform hybrid search (combination of keyword and vector search) in a collection.

  • Parameters:
    • collection_name (required): Name of the collection to search in
    • query_text (required): Text query for hybrid search
    • limit (optional): Maximum number of results to return (default: 5)
    • alpha (optional): Balance between keyword (0) and vector (1) search (default: 0.5)
      • 0.0 = keyword-only search
      • 0.5 = balanced hybrid search
      • 1.0 = vector-only search

🔧 Usage

Note: You need a litle techinical knowledge to run this server.

  • Step 1: Clone this repository
  • Step 2: Initialize your weaviate database
  • Step 3: Add this config to your agentic app
{
  "mcpServers": {
    "weaviate": {
      "args": [
        "--directory",
        "/absolute/path/to/mcp-server-weaviate",
        "run",
        "mcp_server_weaviate",
        "--weaviate-http-port",
        "<weaviate-http-port>",
        "--weaviate-grpc-port",
        "<weaviate-grpc-port>"
      ],
      "command": "uv",
      "disabled": false
    }
  }
}

🧩 Reference

python-sdk

servers

weaviate

squad-explorer

squad-dataset

iteam1/mcp-server-hello

neo4j-contrib/mcp-neo4j-data-modeling

blazickjp/arxiv-mcp-server