LM Studio Keyboard Shortcuts & Commands (22)
17 LM Studio shortcuts and commands.
LM Studio has two faces: a desktop chat app with a small set of keyboard shortcuts, and a local server driven by the lms command-line tool and an OpenAI-compatible HTTP API. The table covers both, because most people start in the app and end up scripting against the server once a model works.
Related Shortcut Pages
Search all shortcuts interactively
Open Shortcut FinderLooking for model parameters? They moved to the new LM Studio prompt commands reference.
All Shortcuts
UI Shortcuts
| Shortcut | Action | Description |
|---|---|---|
| Ctrl + N | New chat | Start a new conversation. |
| Ctrl + Shift + N | New window | Open a new LM Studio window. |
| Ctrl + , | Settings | Open application settings. |
| Ctrl + Enter | Send message | Send the current message. |
| Shift + Enter | New line | Insert a line break without sending. |
| Ctrl + K | Clear chat | Clear the current chat history. |
| Ctrl + F | Find models | Search available models in the browser. |
| Ctrl + D | Model details | View details of the selected model. |
Server & API
| Shortcut | Action | Description |
|---|---|---|
| lms server start | Start server | Start the local LM Studio API server (CLI). |
| lms server stop | Stop server | Stop the running API server. |
| lms server status | Server status | Check whether the server is running. |
| lms ls | List models | List all downloaded models via CLI. |
| lms load <model> | Load model | Load a specific model into memory. |
| lms unload | Unload model | Unload the currently active model. |
| GET /v1/models | List models (API) | OpenAI-compatible endpoint to list loaded models. |
| POST /v1/chat/completions | Chat API | OpenAI-compatible chat completions endpoint. |
| POST /v1/completions | Completion API | OpenAI-compatible text completion endpoint. |
In the app
Ctrl + N starts a new chat with the currently loaded model and Ctrl + Shift + N a new window. Ctrl + Enter sends the message and Shift + Enter adds a line, which is the inverse of the web chat apps and worth noticing on day one. Ctrl + K clears the conversation, Ctrl + F jumps to the model search, and Ctrl + D shows the loaded model's details — quantisation, context length, and the GPU offload in effect. On macOS these are Cmd.
Running a local server
lms server start starts the API server (the app must be installed; the CLI ships with it) and lms server status confirms the port. lms ls lists downloaded models with their sizes, lms load <model> loads one into memory with optional flags for context length and GPU layers, and lms unload frees it. lms server stop shuts the server down. The CLI is what you want in a script or a systemd unit; the app's Developer tab does the same things with switches.
Talking to it
The server speaks the OpenAI wire format, so any client that accepts a custom base URL works unchanged. GET /v1/models lists what is loaded, POST /v1/chat/completions is the chat endpoint that takes a messages array and streams if asked, and POST /v1/completions is the older plain-prompt form. Point an OpenAI SDK at http://localhost:1234/v1 with any string as the API key and the rest of the code stays the same.
Frequently asked questions
Why does Enter add a new line instead of sending?
LM Studio sends on Ctrl + Enter; Shift + Enter inserts a line break. The behaviour of plain Enter can be changed in settings.
How do I start the server without opening the app window?
Run lms server start from a terminal. The lms CLI is installed alongside the app; the first run may ask you to bootstrap it.
Can I use the OpenAI Python library with LM Studio?
Yes. Set base_url to http://localhost:1234/v1 and any api_key value. The chat completions endpoint is compatible.
How do I see which model is loaded?
lms ls shows downloaded models and marks loaded ones; GET /v1/models returns the loaded list from the API.
🤖 Ask AI about LM Studio shortcuts
Open your assistant with this page preloaded as the source — great for follow-up questions like "which of these work in other apps?"