← Back to Weaviate MCP Server

Install Weaviate MCP Server

Pick your client, copy the command, done.

database
Manual setup required. The maintainer's config contains paths only you know - edit the placeholders below before adding it to Claude Code.
1

Prepare the server locally

Run this once before adding it to Claude Code.

git clone https://github.com/sajal2692/mcp-weaviate
cd mcp-weaviate

Then follow the repository README for any remaining dependency or build steps.

2

Register it in Claude Code

claude mcp add mcp-weaviate -- uvx mcp-weaviate --connection-type local --host localhost --port 8080 --grpc-port 50051 --openai-api-key YOUR_OPENAI_API_KEY

Replace any placeholder paths in the command with the real path on your machine.

Environment Variables

Set these before running Weaviate MCP Server.

VariableDescriptionRequired
WEAVIATE_CLUSTER_URLWeaviate Cloud Services URLNo
WEAVIATE_API_KEYAPI key for authenticationNo
OPENAI_API_KEYOpenAI API key for embeddingsNo
COHERE_API_KEYCohere API key for embeddingsNo

Available Tools (7)

Once configured, Weaviate MCP Server gives your AI agent access to:

searchSimplified search interface using hybrid search by default.
querycollection
semantic_searchVector similarity search using embeddings for semantic matching.
querycollection
keyword_searchBM25-based keyword search for exact term matching.
querycollection
hybrid_searchCombined semantic and keyword search with configurable weighting.
querycollectionalpha
list_collectionsList all available collections in your database.
get_schemaGet detailed schema information for specific collections or all collections.
collection_name
get_collection_objectsRetrieve objects from collections with pagination support.
collection_namelimitoffset

Try It Out

After setup, try these prompts with your AI agent:

→List all the collections currently available in my Weaviate database.
→Perform a hybrid search for 'customer support documentation' in the 'knowledge_base' collection.
→Get the schema details for the 'products' collection.
→Retrieve the first 10 objects from the 'user_feedback' collection.
→Check if the 'multi_tenant_data' collection has multi-tenancy enabled.
Prerequisites & system requirements
  • An MCP-compatible client (Claude Code, Cursor, Windsurf, Claude Desktop, or Codex)
  • Python 3.8+ with pip installed

Keep this setup from going cold

Save the docs, env vars, and workflow around Weaviate MCP Server in Conare so Claude Code, Codex, and Cursor remember it next time.

Remember this setup
← Weaviate MCP Server overview Browse database servers →