Weaviate MCP Server (Official)
Official Weaviate MCP server from Weaviate enabling AI agents to interact with Weaviate vector database — performing semantic search and hybrid queries, managing collections and objects, executing near-vector searches, and integrating Weaviate's AI-native vector database into agent-driven RAG, semantic search, and knowledge base workflows.
Score Breakdown
⚙ Agent Friendliness
🔒 Security
TLS enforced for Weaviate Cloud. API key auth. SOC2, GDPR. Official Weaviate MCP. For self-hosted: configure TLS. Protect API key as secret.
⚡ Reliability
Best When
An agent needs vector-based semantic search or RAG over a Weaviate collection — for knowledge base queries, similar document retrieval, or AI-native data access.
Avoid When
You need relational SQL queries, or you're using Pinecone/Qdrant/Chroma — those vector databases have their own MCPs with better coverage.
Use Cases
- • Performing semantic search over vector-embedded content from RAG agents
- • Running hybrid (vector + keyword) searches from search agents
- • Managing vector collections and ingesting documents from data pipeline agents
- • Building knowledge bases for agent memory from AI application agents
- • Querying similar items by vector proximity from recommendation agents
- • Exploring collection schema and vector space from data discovery agents
Not For
- • Teams using Pinecone, Qdrant, or Chroma for vector storage (use those MCPs)
- • Relational data queries (Weaviate is vector-first, not relational)
- • Teams not using embedding-based similarity search
Interface
Authentication
Weaviate Cloud uses API key authentication. Self-hosted supports no-auth (dev) or API key. OpenID Connect (OIDC) for enterprise. Weaviate Cloud connection URL + API key is standard configuration.
Pricing
Weaviate Embedded (local) is free. Weaviate Cloud offers sandbox for development. Production clusters require paid plan. Self-hosted Weaviate is open source and free.
Agent Metadata
Known Gotchas
- ⚠ Vectors must be pre-generated or use Weaviate vectorizer modules — not automatic
- ⚠ Weaviate GraphQL query syntax is specific — not standard GraphQL; agents need Weaviate-specific patterns
- ⚠ Collection schema must define vectorizer and properties before ingesting data
- ⚠ Near-vector vs near-text searches have different behaviors depending on configured modules
- ⚠ Weaviate v3 (Python client) vs v2 have different API patterns
- ⚠ Official Weaviate MCP is newer — verify feature coverage for your use case
Alternatives
Full Evaluation Report
Detailed scoring breakdown, competitive positioning, security analysis, and improvement recommendations for Weaviate MCP Server (Official).
Scores are editorial opinions as of 2026-03-06.