VS Code now supports bring your own key (BYOK), letting developers add language models from providers like Azure, Anthropic, OpenAI, Gemini, Huggingface, and OpenRouter, or run models locally via Ollama or Foundry Local. BYOK models appear in the same Chat model picker used for GitHub Copilot and work without a GitHub account or Copilot subscription. Setup involves opening the Language Models editor, selecting a provider, and entering API credentials. VS Code also supports provider extensions discoverable via the @tag:language-models search. BYOK applies to chat and agent workflows but not inline code completions or embedding-based features. Developers can configure different models for different tasks — fast models for quick queries, reasoning models for complex work, and local models for offline use.
