weaviate-setup

Weaviate Knowledge Store Setup

You are guiding a user through setting up Weaviate as the persistent knowledge store for the video-research MCP server. All 28 tools automatically write results to Weaviate when configured. 8 knowledge tools (knowledge_search, knowledge_related, knowledge_stats, knowledge_fetch, knowledge_ingest, knowledge_schema, knowledge_ask, knowledge_query) enable semantic search and AI-powered Q&A across accumulated research.

Setup Flow

Follow these steps IN ORDER. Use AskUserQuestion for each decision point.

Step 1: Deployment Type

Ask the user which Weaviate deployment they want to use:

AskUserQuestion:
  questions:
    - question: "Which Weaviate deployment will you use?"
      header: "Deployment"
      multiSelect: false
      options:
        - label: "Weaviate Cloud (Recommended)"
          description: "Managed cloud service at console.weaviate.cloud — free tier available, no infrastructure to manage"
        - label: "Local Docker"
          description: "Run Weaviate locally via Docker on port 8080 — full control, no network latency"
        - label: "Custom/Self-hosted"
          description: "Your own Weaviate deployment at a custom URL"

Step 2: Collect Credentials (based on choice)

If Weaviate Cloud:

• Tell the user to go to https://console.weaviate.cloud, create a free cluster, then copy the cluster URL and API key
• Ask them to provide both values

If Local Docker:

• Provide the docker-compose snippet:

services:
  weaviate:
    image: cr.weaviate.io/semitechnologies/weaviate:1.28.4
    ports:
      - "8080:8080"
      - "50051:50051"
    environment:
      QUERY_DEFAULTS_LIMIT: 25
      AUTHENTICATION_ANONYMOUS_ACCESS_ENABLED: "true"
      PERSISTENCE_DATA_PATH: "/var/lib/weaviate"
      DEFAULT_VECTORIZER_MODULE: text2vec-weaviate
      ENABLE_MODULES: text2vec-weaviate

• Note: Our collections use the text2vec-weaviate vectorizer (Weaviate's built-in embedding service). No sidecar container needed.
• The URL will be http://localhost:8080
• No API key needed for local