Twinny

Twinny is an AI coding assistant for developers who want completion and help with their repository inside VS Code, using models on their own hardware or a chosen server.

• Completion and edits: Get suggestions based on surrounding code, request changes in the editor and accept or reject individual parts of the proposed diff.
• Repository chat: Ask questions with context from files, symbols, diagnostics, Git changes and terminal output, plus keyword and vector search across the workspace.
• Reviews and terminal help: Review uncommitted changes, branches or GitHub pull requests, generate commit messages and get proposed commands or fixes for failed commands.
• Shared team models: A self-hosted gateway gives developers access to shared inference servers, tracks usage per person and can pool models running on teammates' computers, switching machines when one disconnects.

Twinny supports Ollama, llama.cpp, LM Studio and OpenAI-compatible servers, as well as hosted providers including OpenAI and Anthropic. Completion and chat can use different backends. With local backends, it works offline, including on air-gapped networks, without an account or telemetry. Requests sent to hosted providers leave your network.

The extension and gateway are open source under MIT. The extension is free, and paid team licensing is available. Shared computers pass prompts through without storing them; gateway content recording is off by default.