Pinecone

Pinecone vector database integration for managing indexes, querying, and rapid prototyping with slash commands and MCP server support.

Play video

Integrate Pinecone vector database directly into your Claude Code workflow. This plugin provides a complete toolkit for managing vector indexes, querying data with natural language, and rapidly prototyping semantic search applications—all without leaving your editor.

The MCP server gives you access to seven powerful tools: list-indexes, describe-index, describe-index-stats, search-records, create-index-for-model, upsert-records, and rerank-documents. Build RAG applications, recommendation systems, and semantic search features with full control over your vector data.

How to use: Run /pinecone:quickstart to generate AGENTS.md files and walk through a Python tutorial. Use /pinecone:query to search your indexes interactively—specify an index name, namespace, and optional reranker. Run /pinecone:help for setup guidance and troubleshooting. You can also ask Claude directly to "list my Pinecone indexes" or "create a new index for my embeddings model."

Requires a Pinecone API key set as PINECONE_API_KEY environment variable before starting Claude Code. Sign up for a free account at pinecone.io to get started.