OmniRoute Lite 🚀

A lightweight, zero-cost AI proxy gateway rewritten in Python and optimized for Debian 12. Consolidate your VSCode AI traffic into a single, 100% free local loopback server.

✨ Features

• Native VSCode Compatibility: Seamless integration with Continue, Cline, and Cursor.
• Zero Operating Cost: Exclusive routing through public endpoints and free tiers.
• Transparent Failover: If a provider goes down or encounters a Rate Limit, traffic is automatically re-routed in real-time without breaking your coding workflow.
• Low Footprint: Runs with a target memory usage of < 50MB RAM.

🛠️ Installation on Debian 12

Prerequisites

Ensure you have Python 3.11+ and the virtual environment manager installed on your system:

sudo apt update
sudo apt install python3-pip python3-venv -y

Environment Setup

1. Clone this repository or copy the project files into a local folder.
2. Build and initialize your isolated virtual environment:

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

🚀 Running the Server

Launch the local gateway using the following module command:

python3 -m src.main

The server will boot and start listening natively on http://127.0.0.1:8000.

💻 VSCode Configuration (Example using Continue Extension)

Inside the config.json file of your Continue extension, add the following model block pointing directly to OmniRoute Lite:

{
  "models": [
    {
      "title": "OmniRoute Lite",
      "provider": "openai",
      "model": "omniroute-free",
      "apiBase": "http://127.0.0.1"
    }
  ]
}

Project Workspace Directory Structure

omniroute-lite/
├── .venv/
├── docs/
│   ├── PLAN.md
│   ├── SPEC.md
│   └── TROUBLESHOOTING.md
├── src/
│   ├── gateway/
│   │   ├── __init__.py
│   │   ├── routes.py
│   │   └── schemas.py
│   ├── router/
│   │   ├── __init__.py
│   │   └── engine.py
│   ├── __init__.py
│   ├── config.py
│   └── main.py
├── tests/
│   └── test_gateway.py
├── .gitignore
├── CHANGELOG.md
├── README.md
└── requirements.txt