This feature is in early access and is not intended for production usage.

The Pinecone MCP server enables AI agents to interact directly with Pinecone’s functionality and documentation via the standardized Model Context Protocol (MCP). Using the MCP server, agents can search Pinecone documentation, manage indexes, upsert data, and query indexes for relevant information.

This page shows you how to configure Cursor and Claude Desktop to connect with the Pinecone MCP server.

​
Tools

The Pinecone MCP server provides the following tools:

  • search-docs: Search the official Pinecone documentation.
  • list-indexes: Lists all Pinecone indexes.
  • describe-index: Describes the configuration of an index.
  • describe-index-stats: Provides statistics about the data in the index, including the number of records and available namespaces.
  • create-index-for-model: Creates a new index that uses an integrated inference model to embed text as vectors.
  • upsert-records: Inserts or updates records in an index with integrated inference.
  • search-records: Searches for records in an index based on a text query, using integrated inference for embedding. Has options for metadata filtering and reranking.
  • cascading-search: Searches for records across multiple indexes, deduplicating and reranking the results.
  • rerank-documents: Reranks a collection of records or text documents using a specialized reranking model.

The Pinecone MCP supports only indexes with integrated embedding. Indexes for vectors you create with external embedding models are not supported.

​
Before you begin

Ensure you have the following:

​
Configure Cursor

1

Add the MCP server

In your project root, create a .cursor/mcp.json file, if it doesn’t exist, and add the following configuration:

{
  "mcpServers": {
    "pinecone": {
      "command": "npx",
      "args": [
        "-y", "@pinecone-database/mcp"
      ],
      "env": {
        "PINECONE_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

Replace YOUR_API_KEY with your Pinecone API key.

2

Check the status

Go to Cursor Settings > MCP. You should see the server and its list of tools.

3

Add Pinecone rules

The Pinecone MCP server works well out-of-the-box. However, you can add explicit rules to ensure the server behaves as expected.

In your project root, create a .cursor/rules/pinecone.mdc file and add the following:

### Tool Usage for Code Generation

- When generating code related to Pinecone, always use the `pinecone` MCP and the `search_docs` tool.

- Perform at least two distinct searches per request using different, relevant questions to ensure comprehensive context is gathered before writing code.

### Error Handling

- If an error occurs while executing Pinecone-related code, immediately invoke the `pinecone` MCP and the `search_docs` tool.

- Search for guidance on the specific error encountered and incorporate any relevant findings into your resolution strategy.

### Syntax and Version Accuracy

- Before writing any code, verify and use the correct syntax for the latest stable version of the Pinecone SDK.

- Prefer official code snippets and examples from documentation over generated or assumed field values.

- Do not fabricate field names, parameter values, or request formats.

### SDK Installation Best Practices

- When providing installation instructions, always reference the current official package name.

- For Pinecone, use `pip install pinecone` not deprecated packages like `pinecone-client`.
4

Test the server

Press Command + i to open the Agent chat. Test the Pinecone MCP server with prompts that required the server to generate Pinceone-compatible code and perform tasks in your Pinecone account.

Generate code:

Write a Python script that creates a dense index with integrated embedding, upserts 20 sentences about dogs, waits 10 seconds, searches the index, and reranks the results.

Perform tasks:

Create a dense index with integrated embedding, upsert 20 sentences about dogs, waits 10 seconds, search the index, and reranks the results.

​
Configure Claude Desktop

1

Add the MCP server

Go to Settings > Developer > Edit Config and add the following configuration:

{
  "mcpServers": {
    "pinecone": {
      "command": "npx",
      "args": [
        "-y", "@pinecone-database/mcp"
      ],
      "env": {
        "PINECONE_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

Replace YOUR_API_KEY with your Pinecone API key.

2

Check the status

Restart Claude Desktop. On the new chat screen, you should see a hammer (MCP) icon appear with the new MCP tools available.

3

Test the server

Test the Pinecone MCP server with prompts that required the server to generate Pinceone-compatible code and perform tasks in your Pinecone account.

Generate code:

Write a Python script that creates a dense index with integrated embedding, upserts 20 sentences about dogs, waits 10 seconds, searches the index, and reranks the results.

Perform tasks:

Create a dense index with integrated embedding, upsert 20 sentences about dogs, waits 10 seconds, search the index, and reranks the results.

​
Configure Claude Code

1

Add the MCP server

Run the following command to add the Pinecone MCP server to your Claude Code instance:

claude mcp add-json pinecone-mcp \
  '{"type": "stdio",
    "command": "npx",
    "args": ["-y", "@pinecone-database/mcp"],
    "env": {"PINECONE_API_KEY": "YOUR_API_KEY"}}'

Replace YOUR_API_KEY with your Pinecone API key.

2

Check the status

Restart Claude Code. Then, run the /mcp command to check the status of the Pinecone MCP. You should see the following:

  > /mcp 
    ⎿  MCP Server Status

      • pinecone-mcp: ✓ connected
3

Test the server

Test the Pinecone MCP server with prompts to Claude Code that require the server to generate Pinceone-compatible code and perform tasks in your Pinecone account.

Generate code:

Write a Python script that creates a dense index with integrated embedding, upserts 20 sentences about dogs, waits 10 seconds, searches the index, and reranks the results.

Perform tasks:

Create a dense index with integrated embedding, upsert 20 sentences about dogs, waits 10 seconds, search the index, and reranks the results.