PlugWiki
Pinecone Developer MCP logo

Pinecone Developer MCP Server: Setup for Claude, Cursor and Other Apps

OfficialLocal
Getnpm@pinecone-database/mcppinecone-io/pinecone-mcp

Lets Claude, Cursor and other assistants search Pinecone docs and create, fill and search your Pinecone indexes.

Pinecone's own developer server gives your coding assistant 12 tools to search the Pinecone docs, set up indexes, add records and run searches against them. It runs on your computer and uses a Pinecone API key from the console. Without a key, only the docs search works. It handles indexes with built-in embedding models and does not support your own vectors or Pinecone Assistants.

Good for

  • Developers building search or retrieval features on Pinecone
  • Teams testing queries and embedding models against real index data
  • Anyone who wants accurate Pinecone answers from the docs inside a coding assistant

Works in

12 should work

Tap an app to see how to set it up and any known problems.

Install Pinecone Developer MCP

Gemini CLI

  1. 1

    Copy this prompt.

    Add the Pinecone Developer MCP server to Gemini CLI. Run this command:
    
    gemini extensions install https://github.com/pinecone-io/pinecone-mcp
    
    Ask me for PINECONE_API_KEY and only put the values in that config, nowhere else.
    
    Then check that it loaded and tell me if I need to sign in or restart.
  2. 2

    Paste it into Gemini CLI. It shows the change and asks before running it.

Requires Node.js 22 or later, PINECONE_API_KEY set in your environment.

The extension route needs your Pinecone key in the PINECONE_API_KEY environment variable, set before you run gemini. The npx route below it is adapted from the generated config and has the same key requirement. Get a key from the Pinecone console; without one only documentation search works.

From the project README. Source.

Tools

  • Searches the Pinecone documentation

    Sends your question to a Pinecone-hosted docs assistant and returns relevant documentation. It works without an API key.

    Try: “Search the Pinecone docs for how metadata filtering works.”

    Network · search-docs

  • Lists your Pinecone indexes

    Returns every index in the project your API key belongs to.

    Try: “List all my Pinecone indexes.”

    Network · list-indexes

  • Shows how one index is set up

    Returns the configuration of an index, such as its model, fields and cloud region.

    Try: “Describe the configuration of my-docs.”

    Network · describe-index

  • Shows record counts for an index

    Returns statistics for an index, including the number of records and the namespaces it holds.

    Try: “What namespaces exist in my-docs and how many records are in each?”

    Network · describe-index-stats

  • Creates a document index

    Creates a new index whose documents have text fields that can be searched by keyword, by meaning through a hosted embedding model, or both.

    Try: “Create a document index called notes with a searchable text field.”

    Write · create-document-index

  • Adds or replaces documents in a document index

    Inserts documents, or replaces any that already have the same ID. The tool is marked as overwriting existing data.

    Try: “Upsert these three documents into the notes index.”

    Write · upsert-documents

  • Searches a document index

    Searches by meaning or by keywords, with optional filters.

    Try: “Search notes for authentication best practices.”

    Network · search-document-index

  • Creates an older-style integrated index

    Creates a legacy integrated index that turns text into vectors with a hosted embedding model you pick.

    Try: “Create an index my-docs with the multilingual-e5-large model.”

    Write · create-index-for-model

  • Adds or replaces records in an index

    Inserts text records into an integrated index, and replaces any with the same ID. The tool is marked as overwriting existing data.

    Try: “Upsert these documents into my-docs.”

    Write · upsert-records

  • Searches records in an index by text

    Searches an integrated index with a text query, with metadata filters and optional reranking.

    Try: “Search my-docs for records related to authentication.”

    Network · search-records

  • Searches several indexes and merges the results

    Searches across multiple indexes, removes duplicates and reranks what is left into one list.

    Try: “Search both my-docs and notes for deployment steps.”

    Network · cascading-search

  • Reorders documents by relevance

    Sends records or text to a reranking model and returns them sorted by how well they match a query.

    Try: “Rerank these five passages for the query "rate limits".”

    Network · rerank-documents

Read from the source code

Permissions

medium risk · 4 things to watch

Medium risk

It can read your indexes and create or overwrite them and their records, but it has no delete tool. Use an API key for a development project and avoid your production data.

What it can do

  • Read your data
  • Create and change things
  • Delete things
  • Read-only mode

Watch out for

  • It can create indexes and overwrite records

    create-document-index and create-index-for-model create indexes. upsert-documents and upsert-records replace any record with the same ID. Use a key for a dev project, not production. github.com

  • The assistant reports its model name to Pinecone

    Every database tool has optional fields where the assistant states its provider and model, and the server passes them to Pinecone, which says it uses them for usage analytics. The assistant is told not to ask you for them. github.com

  • Docs search goes to a Pinecone-hosted service

    search-docs sends your question to a Pinecone-run docs assistant, with or without an API key. github.com

  • No tool deletes an index or records

    The source registers 12 tools and none of them is a delete tool. github.com

Security advisories

No published security advisories as of Nov 1, 2026.

Technical details
Talks toPinecone API (api.pinecone.io). Pinecone-hosted docs assistant (prod-1-data.ke.pinecone.io)
Sign-in and keysPINECONE_API_KEY sits in your MCP client config on your computer and is sent to Pinecone.

Questions

Is the Pinecone MCP server official?

Yes. It is published by Pinecone (the pinecone-io GitHub organization) as the Pinecone Developer MCP Server, so it comes from the maker of the database.

Do I need a Pinecone API key?

Yes for anything beyond docs search. Create a key in the Pinecone console and put it in PINECONE_API_KEY in your config. Without a key only the documentation search works.

What does it need installed?

Node.js 22 or later, with node and npx on your PATH. If the server does not appear, check node --version first.

Can it delete my indexes?

No. None of its 12 tools deletes an index or a record. It can create indexes and overwrite records that share an ID.

Can I use it read-only?

No read-only switch exists. Eight tools only read, and four create indexes or write records. Use a key for a dev project to limit damage.

Does it work with my own vectors or Pinecone Assistant?

No. The README says assistants, standalone embeddings and search with your own vectors are not supported. Pinecone offers a separate Assistant MCP for knowledge bases.

How do I add it to Gemini CLI?

Use the extension command in the Gemini CLI tab of the install box, then set your key in PINECONE_API_KEY. Other apps use the npx config in their tabs.

Information

updated Nov 2, 2026

Publisher
Official
Package
@pinecone-database/mcp
Registry
npm
Transport
stdio
Auth
API key
Version
0.4.0
License
Apache-2.0
Last commit
Oct 29, 2026
Open issues
11

WebsiteGitHubnpmChangelogIssues

Sources

Docs, issue trackers and discussions this page is built from. Issues, quotes and findings above link to their own source.

Show all 5
  1. github.com · github.com
  2. docs.pinecone.io · docs.pinecone.io
  3. github.com · github.com
  4. github.com · github.com
  5. github.com · github.com

Facts from the official MCP registry, GitHub, npm, PyPI and the Docker catalog. Updated Nov 2, 2026.