This MCP server provides a dedicated integration for retrieving information from Pinecone Assistant within the MCP (Model Context Protocol) framework. It acts as a bridge between MCP clients and the Pinecone Assistant API, sending queries to Pinecone Assistant and returning retrieved information to the requester. The server supports retrieving multiple results per query, with the number of results configurable by the client or deployment configuration. It is packaged as a Docker image (pinecone/assistant-mcp) and can be run with environment variables including PINECONE_API_KEY (required) and PINECONE_ASSISTANT_HOST (optional; defaults to https://prod-1-data.ke.pinecone.io). A LOG_LEVEL option is also available for logging. The README provides a Docker run example and a Claude Desktop configuration snippet to integrate the MCP server into the Claude Desktop workflow. For those who prefer building from source, Rust is used; after cloning the repo, run cargo build --release to produce the assistant-mcp binary at target/release/assistant-mcp. There is also guidance for testing with the inspector, and prerequisites include a Pinecone API key and a Pinecone Assistant host obtained from the Pinecone Console.