Ollama Wake-on-LAN Proxy

ollama-wol-proxy is a Caddy-based reverse proxy with Wake-on-LAN (WoL) support for an Ollama inference server. It automatically wakes up the server when a request is made, ensuring seamless access while saving power by allowing the server to sleep when idle.

Features

• Wake-on-LAN Integration: Automatically sends a WoL magic packet to wake the server if it's offline.
• Reverse Proxy: Forwards requests to the Ollama server once it's online.
• Customizable: Easily configure MAC and IP addresses using environment variables.
• Dockerized: Ready-to-use Docker image with Caddy and WoL plugin.

Prerequisites

• Docker and Docker Compose installed.
• A server with Wake-on-LAN enabled.

Setup

1. Clone this repository:

   git clone https://github.com/your-username/ollama-wol-proxy.git
   cd ollama-wol-proxy

2. Create a .env file in the project root with the following content:

   WOL_MAC=3C:7C:3F:D6:00:CB
   SERVER_IP=10.0.0.155

3. Build and run the Docker container:

   docker-compose up -d

Configuration

• WOL_MAC: The MAC address of your server.
• SERVER_IP: The IP address of your Ollama server.

License

This project is licensed under the MIT License. See the LICENSE file for details.