This repository contains a minimal LiteLLM proxy setup for routing requests to GitHub Copilot models.

• Runs the LiteLLM proxy in Docker.
• Loads model configuration from config.yaml.
• Persists GitHub Copilot token data on the host.
• Provides a compatibility test script for proxy endpoints.

• run-docker.sh: Starts the LiteLLM proxy container with the required mounts and env vars.
• config.yaml: Model and header configuration used by the proxy.
• compatibility_check.py: Runs sequential endpoint tests and writes compatibility_table.md.
• compatibility_table.md: Generated compatibility matrix for the configured models.

1. Start the proxy:

   sh run-docker.sh

2. Run compatibility checks:

   python3 compatibility_check.py

• GitHub Copilot headers are defined once and reused across models via YAML anchors.
• The proxy listens on http://localhost:4000 by default.
• Token files are stored under the host directory mounted by run-docker.sh.

This is a lightweight, local setup intended for development and testing of the proxy configuration and model compatibility.