What it is
LM Studio is a desktop app for running AI language models locally. You browse and download open models such as Llama, Qwen, Gemma or Mistral from inside the app, then chat with them in a window, with no account and no data sent to a cloud service.
It can also serve a loaded model over a local API that’s compatible with OpenAI’s, so other apps and scripts on your computer can use it in place of an online provider.
Why it’s here
- Local and private. Models run on your hardware, and your prompts stay on your computer.
- A model catalog built in. It shows which versions of a model will fit your memory.
- A local server. Point developer tools at it to test against a model without paying per request.
On CoreDesktop
LM Studio is installed from Flathub. Open it from the app menu. Larger models need plenty of memory and run best with a capable GPU. update keeps LM Studio current with the rest of the Flatpaks.