Twinny

Twinny is a VS Code extension that provides Fill-in-the-Middle (FIM) code completion, AI-powered chat for code discussion (including explanations, test generation, and refactoring suggestions), code generation, and optimization. It can connect to various large language models, especially local ones like those served via Ollama, giving developers intelligent assistance as they code. Its key strengths are full open-source transparency and a strong emphasis on privacy: code never leaves the user’s control and can be processed entirely locally. Users can choose different model providers and customize prompt templates to fit their workflows. There is no subscription cost, making it attractive for privacy-focused or budget-conscious teams. However, the setup requires self-deploying and configuring a model service, which demands technical know-how. The completion quality and speed depend heavily on the chosen model and the hardware running it, and may fall short of commercial cloud-based assistants. As a community project, it might have less polished documentation or slower support.