beemcp

BeeMCP is an unofficial Model Context Protocol (MCP) server that bridges Bee.computer data to language models and AI agents. It exposes your Bee data — including conversations, facts, to-dos, and location history — in a structured way that LLMs can query and optionally modify. By running the server with an MCP client (such as Claude or Zed) configured to start the service, you enable the AI to list conversations, retrieve full conversation details, read and manage user facts, and interact with your reminders and location history through the provided tool calls. The server requires a Bee API key and uses it to securely access your Bee data; you’ll typically provide this key via an environment variable (BEE_API_TOKEN) in your MCP client configuration or in a local .env file when running the server directly.

With BeeMCP, you can use the Available Tools (Conversations, Facts) to perform operations like listing all conversations, fetching a specific conversation by ID, listing user facts, getting or updating a fact, recording new facts, and deleting facts. These actions allow an AI assistant to reason about your past interactions and stored preferences, then present summaries, reminders, or context-aware insights to you or an external agent. The server is designed to be run locally and integrated into your existing MCP client workflow, making Bee data accessible to an AI assistant while keeping control of the API key in your environment.

Prerequisites:

• Python 3.10 or higher
• Bee.computer API key (BEE_API_TOKEN)
• Access to install Python packages (pip) or uv (recommended)

Installation options:

1. Using uv (recommended)

• If you already have uv installed, you can run beemcp directly without a separate installation:

# Ensure uvx is available (via uv)
# Run the server (you will be prompted or required to provide BEE_API_TOKEN via env)
uvx beemcp

2. Using pip

• Install beemcp from PyPI:

pip install beemcp

• Run the server as a Python module:

python -m beemcp.beemcp

• Alternatively, if the entry point was added to PATH during installation, you can run:

beemcp

Configuration tips:

• Create a .env file in the directory where you run the server and add:

BEE_API_TOKEN="your_actual_bee_api_key_here"

• Or export the key in your shell before starting the server:

export BEE_API_TOKEN="your_actual_bee_api_key_here"
uvx beemcp

Notes and tips:

• The server requires BEE_API_TOKEN to access Bee data; without it, the server will error out. Store this key securely.
• When configuring clients (Claude or Zed), pass the BEE_API_TOKEN via environment variables in the MCP server settings so the AI can authenticate with Bee.
• Using uvx is the recommended deployment for performance and simplicity. If you install via pip, you can still run the Python module as shown.
• Ensure your Bee API key permissions align with the operations you intend to perform (read vs. write access for facts, conversations, and reminders).
• If you encounter connectivity or authentication errors, verify that the API key is correct, not expired, and that your environment variables are visible to the process launching the MCP server.
• When integrating with an MCP client, verify the correct command and arguments in the client’s settings (for uvx, beemcp; for Python, -m beemcp.beemcp).