LM Studio

The polished desktop app for running local LLMs: download a model from Hugging Face, chat in a native UI, and serve an OpenAI-compatible API on localhost — no terminal required.

Product preview

Overview

LM Studio is the local-LLM app you hand to someone who will not run Ollama in a terminal: a store-like window, a model browser, a chat pane, and a one-click local server that speaks OpenAI so Cursor or a Python script can point at localhost. It is closed-source, free for personal use, and the UI is why it spread beyond the GGUF crowd.

Jan and Ollama remain the open-source alternatives; Studio wins on discoverability and GPU sliders that make a 13B model fit. Hardware still caps quality, and the license is not for a company fleet without reading it. For a private ChatGPT on a laptop, it is the most approachable front door.

Pros / Cons

Pros

Native UI, no terminal tax
One-click local OpenAI server
Hugging Face browsing inside the app
GPU sliders that make mid-size models fit

Cons

Closed source, read the license for work use
Quality still capped by the machine
Not a headless daemon for servers

Who it's for

Private ChatGPT on a laptop with a UI

Local OpenAI endpoint for IDE plugins

Trying Hugging Face models without a cloud bill

Handing local LLMs to non-terminal users

Bottom line

The friendliest door into local LLMs — use it when Ollama's terminal is the reason people bounce.

Key Features

Native apps for macOS, Windows and Linux

Hugging Face model browser

Local OpenAI-compatible server

GPU offload controls

Chat UI with presets and system prompts

Compare LM Studio with alternatives

LM Studio vs Hugging Face LM Studio vs Replicate LM Studio vs Ollama 4 products · AI Platforms

Frequently asked questions

What is LM Studio?

LM Studio is a AI Platforms tool developed by LM Studio. The company was founded in 2023 and is headquartered in Canada. The polished desktop app for running local LLMs: download a model from Hugging Face, chat in a native UI, and serve an OpenAI-compatible API on localhost — no terminal required.