MCP-Pinecone-Notes

An MCP server for managing notes using Pinecone vector database for semantic search capabilities.

Installation

You can install mcp-pinecone-notes from PyPI:

pip install mcp-pinecone-notes

Or install from source:

git clone https://github.com/yourusername/mcp-pinecone-notes.git
cd mcp-pinecone-notes
pip install -e .

Development Setup

For development, first install the package with development dependencies:

pip install -e .[dev]

This will install all required development tools:

• ruff: For linting
• black: For code formatting
• mypy: For type checking
• pytest: For running tests

To run the same lint checks as the CI pipeline:

# Run linting
ruff check .

# Check code formatting
black --check .

# Run type checking
mypy .

# Run tests
pytest

Configuration

The server can be configured through command line arguments or environment variables. Command line arguments take precedence over environment variables.

Command Line Arguments

• --pinecone-api-key: Your Pinecone API key
• --pinecone-host: Host URL for your Pinecone index

Example:

mcp-pinecone-notes --pinecone-api-key YOUR_KEY --pinecone-host YOUR_HOST

With uvx:

uvx run mcp-pinecone-notes -- --pinecone-api-key YOUR_KEY --pinecone-host YOUR_HOST

Environment Variables

The following environment variables are supported:

• PINECONE_API_KEY: Your Pinecone API key
• PINECONE_INDEX: Name of your Pinecone index
• PINECONE_HOST: Host URL for your Pinecone index
• PINECONE_NAMESPACE: (Optional) Namespace for your notes, defaults to "default"

You can set these in a .env file or export them in your shell.

Usage

Basic Usage

Run the MCP server:

mcp-pinecone-notes

Development

For development, you can run directly with:

uv run src/mcp_pinecone_notes/main.py

Use the MCP inspector for testing:

npx @modelcontextprotocol/inspector uvx run mcp-pinecone-notes
npx @modelcontextprotocol/inspector uv run mcp-pinecone-notes -- --pinecone-api-key YOUR_KEY --pinecone-host YOUR_HOST

Docker

Build the Docker image:

docker build -t mcp-pinecone-notes .

Run the container:

docker run \
  -e PINECONE_API_KEY=<key> \
  -e PINECONE_INDEX=<index> \
  -e PINECONE_HOST=<host> \
  -e PINECONE_NAMESPACE=<namespace> \  # Optional, defaults to "default"
  mcp-pinecone-notes

Note: The PINECONE_NAMESPACE environment variable is optional and will default to "default" if not specified.

MCP Tools

add_note

Add a note to storage with semantic embedding:

{
  "name": "add_note",
  "arguments": {
    "note": "Your note text here"
  }
}

The note will be embedded using the nomic-embed-text-v1.5 model and stored in Pinecone with a unique ID and timestamp.

Integration in Claude

{
  "mcpServers": {
    "pinecone": {
      "command": "mcp-pinecone-notes",
      "args": []
    }
  }
}

## Update Agent

```bash
cp -r ~/git/mcp-pinecone-notes/agent/ ~/git/github/llm-functions/agents/notes/