What is LM Studio?

LM Studio is a desktop app from Element Labs for downloading and running open models on your own computer, with llama.cpp and Apple's MLX as its engines. Its newest release centers on Bionic, an agent that creates documents, slides, PDFs, and code using local models or open models hosted in LM Studio's own US cloud.

For developers, LM Studio runs a local server with OpenAI- and Anthropic-compatible endpoints alongside its own REST API, plus TypeScript and Python SDKs and the lms CLI. A headless daemon called llmster runs the same core on servers or in CI without the GUI, and LM Link lets one device use models loaded on another, for up to five devices on the free plan.

Running models locally is free, and the paid plans add usage of US-hosted open models with zero data retention. Team subscriptions aren't available yet, only centralized billing for credits. LM Studio is a runtime and agent rather than a production platform, so routing across providers, evals, and production tracing come from other tools.

Key features

Core capabilities this platform advertises.

Desktop app for browsing, downloading, and running open models
llama.cpp and MLX engines
Local server with OpenAI- and Anthropic-compatible endpoints
TypeScript and Python SDKs and the lms CLI
llmster headless daemon for servers and CI

Strengths and tradeoffs

What this tool does well, and the limitations to keep in mind.

Pros

Free to run local models
MLX engine for Apple Silicon alongside llama.cpp
Local server compatible with both the OpenAI and Anthropic APIs

Cons

No routing across providers or automatic failover
No built-in evals or production tracing

Plans & pricing

What's included in each plan, and how the tiers compare.

Free

Bionic agent, local models on llama.cpp and MLX, offline voice transcription, LM Link for up to 5 devices, limited web search

Bionic+

$20

Per month

Everything in Free, plus US-hosted open models like Kimi K3, GLM 5.3, and DeepSeek V4 Flash, discounted bulk tokens, web search and page extraction

Pro

$100

Per month

Common use cases

Developers and individuals who want to run open models on their own computer through a desktop app or a local API

Running open models privately without the command line
Testing local models against an OpenAI-compatible API
Running models on a home server and using them from a laptop

Using LM Studio with Respan

LM Studio runs open models on your machine. When an app built on those models moves to production, Respan routes across 1,000+ hosted models, including open models from providers like Together AI, Fireworks, and Novita, with tracing, spend limits, and evals on every call.