LM Studio CLI
LM Studio is a desktop app for running open models locally, with a chat, a local server in OpenAI and Anthropic formats, and MCP support.
Extended access
Why this level
- A closed-source app: you cannot review its code before installing
- The local server accepts requests without a key by default; expose it to a network only on purpose
Install
- Works in
Manual install
https://lmstudio.ai/downloadDownload the installer for your system from the official site.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- A closed-source app: you cannot review its code before installing
- The local server accepts requests without a key by default; expose it to a network only on purpose
This is third-party code. Review the repository files before installing.
What it does
LM Studio finds and downloads open models through the app window, no command line needed, and runs them on your machine. It has a chat, chat with documents and model management. The Developer tab or `lms server start --port 1234` starts a local server. It accepts OpenAI-format requests (`/v1/...`, including `/v1/responses`) and Anthropic-format ones (`/v1/messages`), so Claude Code, Codex and OpenCode can connect to it. You can add MCP servers to the app, and the local model will call their tools.
Who it is for. People who want to run a local model through an app window instead of the command line, and developers who need a local server for agents.
Good fit when
- You want to try an open model without terminal setup
- You need to chat with documents offline
- An agent needs a local model with MCP tools
Not a fit when
- You need open source: LM Studio is closed source
- You have an Intel Mac
Example request
Start the LM Studio server on port 1234 and connect Codex to it with codex --ossLimitations
On macOS it needs Apple Silicon and macOS 14 or later. On Windows it needs an x64 CPU with AVX2 or ARM, 16 GB of RAM and at least 4 GB of video memory.
The LM Studio terms allow personal and internal business use; redistribution is not allowed.
For Claude Code the docs call for a context of 25K tokens or more.
How to disable. Stop the server with `lms server stop` or in the Developer tab, and delete unused models in the app's models section.
FAQ
Is LM Studio free?
Yes, the Free plan at $0 includes running local models. The paid Bionic+ and Pro plans add cloud features.
How do I connect LM Studio to Claude Code?
Start the server on port 1234, set ANTHROPIC_BASE_URL=http://localhost:1234 and ANTHROPIC_AUTH_TOKEN=lmstudio, then run `claude --model openai/gpt-oss-20b`.
How do I connect LM Studio to Codex?
Run `lms server start --port 1234` and then `codex --oss`. By default Codex picks openai/gpt-oss-20b.
Does LM Studio run on Intel Macs?
No. On macOS it requires Apple Silicon and macOS 14 or later.
How is LM Studio different from Ollama?
LM Studio is driven through an app window and is not open source. Ollama is an MIT-licensed open project aimed first at the command line.
Discussion
No questions or reviews yet.
Related
CLI toolsA CLI and Python library for OpenAI, Claude, Gemini, local and other models: prompts, chats, tools, embeddings and a SQLite log
A CLI to run LLM prompts and act as an MCP client from the terminal, with any MCP server and any model provider
Ollama runs open models on your own machine and serves them to agents through a local API compatible with OpenAI and Anthropic.
An MCP server that lets Claude Code, Codex and other agents call Gemini, GPT, Grok and local models for review, debugging and consensus
A terminal MCP client for local Ollama models: MCP tools, prompts and resources, an agent mode and model switching
A skill that gives text-only models sight: an image by path or from a paste becomes structured JSON with transcribed text, layout and semantics
README in short
The docs cover system requirements, the local server in OpenAI and Anthropic formats, MCP in the app, the lms CLI and integrations with Claude Code, Codex and OpenCode.