Live Translate can now use models from your local LM Studio (or Bionic) library. Turn it on in Settings → LM Studio and your downloaded models appear in the list.
- The port is detected automatically — nothing to type in
- Models that can read images are marked in the list
- Switching model unloads the old one and loads the new one for you
- Talks to 127.0.0.1 only — nothing leaves your computer
Off by default. If you don't use LM Studio nothing changes; the built-in engine works exactly as before.
Requires: LM Studio running with its local server on, and at least one model already downloaded.
Fixed: Live Translate no longer terminates llama-server or python processes belonging to other applications.