Versions:
copilocal, published by Gary Lumsden and currently at version 0.1.3 with two versions released, is a developer tool in the command-line interface and AI tooling category that simplifies running GitHub Copilot CLI against locally hosted large language models. Its purpose is to let developers pick a local LLM from a terminal user interface (TUI) and launch GitHub Copilot CLI against it, bridging the gap between local model providers and Copilot's bring-your-own-key (BYOK) configuration. The software discovers the models a user already has installed across three supported providers — Ollama, Foundry Local, and LM Studio — and presents them in an arrow-key navigable terminal UI for selection. Once a model is chosen, copilocal ensures that the selected provider's OpenAI-compatible server is running, then launches GitHub Copilot CLI with the correct BYOK environment variables set only on the Copilot child process, leaving the rest of the user's shell environment untouched. This scoped environment handling makes it a practical use case for developers who want to experiment with local models in Copilot CLI workflows without manually configuring providers or polluting global environment settings. Additional use cases include quickly switching between locally available models across multiple providers from a single interface and reducing setup friction for offline or privacy-conscious AI-assisted development. If a provider is not installed, copilocal can install it and provides links to each tool's documentation, lowering the barrier to getting started. copilocal is an independent, unofficial tool and is not affiliated with, endorsed by, or sponsored by GitHub, Microsoft, Ollama, or LM Studio; it launches the GitHub Copilot CLI, which users install separately. The tool therefore functions as a lightweight launcher and configuration helper rather than a replacement for any of the underlying components it orchestrates.
Tags: