pinecone:query

Query Pinecone integrated indexes with natural language text via the Pinecone MCP server.

68|12|Updated Dec 17, 2025
One-click install
npx skills add https://github.com/pinecone-io/pinecone-claude-code-plugin --skill pinecone-query-pinecone-io
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: pinecone:query
Source: https://github.com/pinecone-io/pinecone-claude-code-plugin/tree/main/skills/query
Command: npx skills add https://github.com/pinecone-io/pinecone-claude-code-plugin --skill pinecone-query-pinecone-io

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Searching a Pinecone vector index normally requires embedding text into vectors and writing API calls. This Skill lets you run semantic searches against Pinecone integrated indexes directly with plain text queries through the Pinecone MCP server, without writing embedding or query code. ## Core Features & Use Cases - Natural Language Search: Send text queries to integrated indexes and let the MCP server handle embedding and retrieval automatically. - Interactive Argument Resolution: When index, namespace, or query arguments are missing, the Skill lists available indexes and namespaces and prompts you to choose. - Reranking Support: Optionally apply a reranking model to improve result relevance, with results formatted into a readable table. - Use Case: You have a documentation index built with the multilingual-e5-large embedding model. Ask a question in plain language, pick the namespace, and get ranked records with IDs, scores, and metadata. ## Quick Start Ask the assistant to search your Pinecone index for a phrase, for example: query my docs index for "how to configure vector namespaces" with topK set to 5.

Frequently Asked Questions about pinecone:query

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I query a Pinecone index with a text query?

Use the Pinecone MCP search-records tool with your query text, index name, and optional namespace. The MCP server converts the text into embeddings automatically and returns matching records with scores and metadata.

What is a Pinecone integrated index?

An integrated index is a Pinecone index configured with a built-in hosted embedding model such as multilingual-e5-large, llama-text-embed-v2, or pinecone-sparse-english-v0. Pinecone embeds your text for you, so you can search with plain text instead of precomputed vectors.

Can I query a standard Pinecone index with text using the MCP server?

No. The Pinecone MCP server currently only supports integrated indexes with hosted Pinecone embedding models. For standard indexes, custom vector queries, or third-party embedding models like OpenAI or Cohere, use the pinecone:cli skill instead.

Why does the Pinecone query fail with an access error?

The PINECONE_API_KEY environment variable is missing or invalid. Set it with export PINECONE_API_KEY="your-key" in your terminal, then restart your IDE or agent session so the MCP server can authenticate.

How do I rerank Pinecone search results?

Pass a reranker argument specifying the reranking model when running the query. The MCP server applies the reranking model to the returned documents to improve relevance ordering before displaying results.