Skip to main content
The official Pinecone plugin for Cursor adds Pinecone skills, the Pinecone MCP server, and slash commands to your editor. Use natural language to manage indexes, query data, build RAG applications, and create document Q&A assistants, with up-to-date Pinecone API knowledge.

​
Features

  • Built-in skills cover index management, semantic search, full-text search, assistant creation, and more.
  • The bundled MCP server (@pinecone-database/mcp) runs Pinecone operations directly from Cursor Agent.
  • Slash commands like /pinecone-quickstart and /pinecone-query give you quick access to skills.
  • Cursor Agent invokes the right skill automatically based on your conversation.

​
Prerequisites

  • A Pinecone API key
  • Cursor installed
  • Node.js v18+ (required for the bundled MCP server)
  • uv installed (required to run the bundled Python scripts, including the quickstart skill)
  • Pinecone CLI installed (optional, enables the pinecone-cli skill)

​
Installation

1

Set your API key

Add your Pinecone API key to a .env file at your workspace root:
Cursor loads this file into the MCP server via its envFile field, so you don’t need to export the key in your shell.
2

Install the plugin

In Cursor chat, run:
Or install directly from the Cursor Marketplace.
3

Verify the installation

Open Cursor Agent chat and run /pinecone-help to confirm the skills are loaded. You can also check:
  • Skills are listed under Agent Decides in Cursor Settings > Rules.
  • The MCP server is listed in Cursor Settings > Features > Model Context Protocol.

​
Available skills

​
MCP tools

The plugin includes the Pinecone MCP server, which provides the following tools:
  • search-docs: Search the official Pinecone documentation.
  • list-indexes: List all available Pinecone indexes.
  • describe-index: Get index configuration and namespaces.
  • describe-index-stats: Get record counts and namespace statistics.
  • create-index-for-model: Create a new index with integrated embeddings.
  • upsert-records: Insert or update records in an index.
  • search-records: Search records with optional metadata filtering and reranking.
  • cascading-search: Search across multiple indexes with deduplication and reranking.
  • rerank-documents: Rerank documents using a specified reranking model.
For full MCP server documentation, see Use the Pinecone MCP server.

​
Resources