Pinecone Gemini CLI Extension

The official Pinecone extension for the Gemini CLI. Manage vector indexes, perform semantic search, and build document Q&A assistants directly from your terminal using AI-powered Agent Skills and MCP tools.

Quick Start

# 1. Install the extension
gemini extensions install https://github.com/pinecone-io/gemini-cli-extension

# 2. Set your API key (get one free at https://app.pinecone.io)
export PINECONE_API_KEY="your-api-key"

# 3. Restart Gemini CLI, then ask:
#    "Use the quickstart skill to help me get started with Pinecone"

Note: If you hit API key errors, exit Gemini CLI, run export PINECONE_API_KEY="your-key" in your terminal, and start Gemini CLI again. The CLI only reads environment variables at launch.

Installation

1. Prerequisites

• Pinecone Account: Create a free account at app.pinecone.io.
• Pinecone API Key: Generate a key in the Pinecone console.
• uv (Recommended): To run the Python scripts bundled with the skills, install uv.
• Pinecone CLI (Optional): For advanced vector management, install the CLI:

  brew tap pinecone-io/tap
  brew install pinecone-io/tap/pinecone

2. Add the Extension

gemini extensions install https://github.com/pinecone-io/gemini-cli-extension

3. Configure the API Key

You must set your Pinecone API key in two places to unlock all features:

1. MCP Configuration: Add your key to the PINECONE_API_KEY setting in your Gemini CLI or MCP server configuration.
2. Environment Variable: Export the key in your terminal so the bundled skill scripts can access it:

   export PINECONE_API_KEY="your-api-key"

Usage

Agent Skills

This extension includes specialized Agent Skills that are automatically invoked based on your conversation. You don't need to memorize commands; just state your intent:

• "Show me the Pinecone quickstart"
• "Create a document assistant from these files"
• "Search my 'products' index for 'green shoes'"
• "How do I manage namespaces with the Pinecone CLI?"

Note on Auto-Invocation: If you find the agent isn't picking up a specific skill, you can explicitly ask for it by name: "Use the quickstart skill to help me get started."

Available Skills

Skill	Best Used For...
quickstart	Getting started with vector databases or document assistants.
query	Semantic text search on integrated indexes via MCP.
assistant	RAG-based document Q&A with citations.
cli	Learning and running pc terminal commands.
mcp	Reference for all available Pinecone MCP tools.
pinecone-docs	Looking up official documentation and API references.
help	Overview of all setup requirements and skills.

MCP Tools

The extension also provides direct access to Pinecone MCP tools for low-level operations like list-indexes, describe-index, and search-records.

Troubleshooting

• API Key Errors: If a skill script fails, ensure you have exported PINECONE_API_KEY in your current terminal session.
• Skill Activation: If the agent is improvising instead of using a skill, try saying "Activate the [skill-name] skill".
• Documentation: Use the pinecone-docs skill or visit docs.pinecone.io for help.

Contributing

We welcome contributions! Please see CONTRIBUTING.md for details on how to report issues or submit improvements.

License

MIT