twinny

Start here

Quick start Install the extension and connect it to a local model server in a few minutes

Choosing a model server Local, hosted or another of your devices, and what hardware runs which models

Providers Ollama, LM Studio, llama.cpp, any OpenAI-compatible server, or a hosted API

Troubleshooting What each error means and how to fix it

Teams One gateway on your hardware for the whole team: keys, usage, policy, and pull requests reviewed by your own models

What's new What each release added, with links to the pages that explain it

What twinny does

Code completion

Fill-in-the-middle suggestions as you type, with context from open files, recent edits and IntelliSense. Press Tab to accept.

Inline edit

Press Ctrl+I, describe a change, and review it as a diff in the editor. Accept, reject, or refine.

Chat

Ask about your code with files, symbols, problems, git changes and terminal output attached with @.

Agent mode

Switch it on and the chat model reads, searches and edits files and runs commands, showing every step. Commands ask before they run unless you allow them.

Workspace index

Embed the workspace so chat can find the code a question is about, ranked by a local reranker.

Code review and commits

Review the working tree, a branch or a GitHub pull request, and write commit messages.

Terminal

Write a command from a description, or fix the last error with the output attached.

Devices

Use the GPU in another of your computers over an encrypted peer-to-peer link. No account, no relay.

Private by default

Everything runs against a server you control. No telemetry, no sign-in, MIT licensed.