Pick a local LLM (Ollama / Foundry Local / LM Studio) from a TUI and launch GitHub Copilot CLI against it.