Pinecone MCP Server

About

## Pinecone MCP Server: Managed Vector Database The **Pinecone MCP Server** integrates Pinecone's fully managed vector database into Google Antigravity. This platform provides blazing-fast similarity search at any scale, with zero operational overhead for building production AI applications. ### Why Pinecone MCP? Pinecone leads in managed vector search: - **Fully Managed**: No infrastructure to maintain - **Blazing Fast**: Sub-100ms queries at any scale - **Hybrid Search**: Combine vectors with metadata filters - **Serverless**: Pay only for what you use - **Enterprise Ready**: SOC2, HIPAA compliant ### Key Features #### 1. Index Management ```python from pinecone import Pinecone pc = Pinecone(api_key="your-key") # Create serverless index pc.create_index( name="semantic-search", dimension=1536, metric="cosine", spec=ServerlessSpec(cloud="aws", region="us-east-1") ) index = pc.Index("semantic-search") ``` #### 2. Vector Operations ```python # Upsert vectors index.upsert( vectors=[ {"id": "doc1", "values": embedding1, "metadata": {"category": "tech"}}, {"id": "doc2", "values": embedding2, "metadata": {"category": "science"}} ], namespace="articles" ) # Query with filters results = index.query( vector=query_embedding, top_k=10, filter={"category": {"$eq": "tech"}}, include_metadata=True ) ``` #### 3. Namespaces ```python # Organize data with namespaces index.upsert(vectors=user_docs, namespace="user-123") index.upsert(vectors=org_docs, namespace="org-456") # Query specific namespace results = index.query( vector=query, namespace="user-123", top_k=5 ) ``` ### Configuration ```json { "mcpServers": { "pinecone": { "command": "npx", "args": ["-y", "@anthropic/mcp-pinecone"], "env": { "PINECONE_API_KEY": "your-api-key", "PINECONE_ENVIRONMENT": "us-east-1" } } } } ``` ### Use Cases **Semantic Search**: Build search that understands meaning across millions of documents. **RAG Applications**: Store embeddings for retrieval-augmented generation systems. **Recommendations**: Find similar products, content, or users at scale. The Pinecone MCP Server brings managed vector search to Antigravity.

Installation

{
  "mcpServers": {
    "pinecone": {
      "mcpServers": {
        "pinecone": {
          "env": {
            "PINECONE_API_KEY": "YOUR_PINECONE_API_KEY",
            "PINECONE_ENVIRONMENT": "YOUR_ENVIRONMENT"
          },
          "args": [
            "-y",
            "pinecone-mcp"
          ],
          "command": "npx"
        }
      }
    }
  }
}

How to Use

1. 11. Get API key from Pinecone console
2. 22. Set environment (e.g., us-east-1-aws)
3. 33. Sub-50ms query latency