Getting Started
Requirements
- Python 3.10+
- Git
- Linux server with systemd (tested on Ubuntu)
- GPU optional but recommended
Installation
Clone the repo and run the setup script. It creates a virtual environment, installs dependencies, and registers a systemd service.
git clone https://github.com/bobcowher/beekeeper.gitcd beekeeperbash setup.shThe setup script will:
- Detect your Python version (3.12, 3.11, 3.10, or python3)
- Create a venv and install dependencies
- Generate and install a systemd service file (requires sudo)
- Enable and start the service
Once complete, BeekeeperML is running on port 5000. Open http://your-server:5000 in a browser.
Managing the Service
# Check statussudo systemctl status beekeeper
# View logsjournalctl -u beekeeper -f
# Restartsudo systemctl restart beekeeper
# Stopsudo systemctl stop beekeeperDevelopment Mode
To run without systemd for development or testing:
cd beekeepersource venv/bin/activatepython app.pyThis runs Flask’s development server on port 5000 with auto-reload.
Connect an AI Agent (MCP)
Install the MCP server:
pip install beekeeper-mcpRegister it with Claude Code:
claude mcp add beekeeper -s user \ -e BEEKEEPER_HOST=http://your-server:5000 \ -- beekeeper-mcpIf auth is enabled on your server, add -e BEEKEEPER_API_KEY=your-key before --.
See MCP Server for full setup details.
Current Limitations
- Git auth — Private repositories work over SSH only. Use an SSH-form Git URL (
git@github.com:you/repo.git) and add BeekeeperML’s public key to GitHub — see SSH Key. HTTPS credentials are not supported. - HTTPS — For HTTPS, put BeekeeperML behind a reverse proxy. It’s not ready to handle anything sensitive without one.
- Multi-server — Single-server product. A central Hive managing multiple workers is planned for the future.
Authentication (login, sessions, API keys, user management) is implemented and available in the admin panel. It’s off by default for home lab use — enable it if your server is exposed beyond your local network.