Pinecone, the managed vector database company, publishes and maintains this server for its customers.

An official MCP server that lets an AI assistant search Pinecone documentation, create indexes with integrated inference, and upsert, search, and rerank vector records.

Runs locally via npx or uvx, or is configured as a remote MCP endpoint tied to a Pinecone account.

Released by Pinecone in 2025 as part of its push to support AI coding assistants and agent workflows.

Lets agents build and query retrieval augmented generation pipelines directly against a production vector index without custom glue code.

Requires an MCP client and a Pinecone API key with access to the target project and indexes.

Official Pinecone repository, Apache 2.0 license, can create indexes and write or delete vector records so it can modify data, available both as a local process and hosted.