Join waitlist
Back to products
LM Studio

LM Studio

Discover, download, and run local LLMs (incl. DeepSeek R1)

Website lmstudio.ai
Overview

What it is

🤖 • Run LLMs on your laptop, entirely offline 📚 • Chat with your local documents 👾 • Use models through the in-app Chat UI or an OpenAI compatible local server

Intent

I need it when

Connect external tools and services to local AI models via standardized protocols

LM Studio supports Model Context Protocol (MCP) servers, allowing models to call external tools, APIs, and services. Users can build agentic workflows that combine local inference with remote integrations while maintaining control over model execution.

Chat with documents and perform retrieval-augmented generation entirely offline

LM Studio's chat interface supports document attachment for offline RAG workflows. Users can analyze PDFs, text files, and other documents using local models without sending data to external services.

Deploy AI inference on servers and CI/CD pipelines without GUI overhead

llmster (LM Studio's headless daemon) runs on Linux servers and cloud instances without a desktop interface. Users can deploy, manage, and serve local models programmatically via CLI and REST API for production workloads.

Run large language models privately on personal hardware without cloud dependencies

LM Studio enables users to download and execute open-source LLMs (Llama, Qwen, DeepSeek, Gemma) locally on Mac, Windows, or Linux. This preserves data privacy, eliminates cloud costs, and allows offline operation after initial model download.

Build AI applications with local models using familiar APIs and SDKs

LM Studio provides TypeScript (lmstudio-js), Python (lmstudio-python), and REST APIs with OpenAI and Anthropic compatibility. Developers can integrate local models into applications, scripts, and agents without vendor lock-in or external API calls.

Drop

Not a fit when

  • User requires cloud-hosted AI without local infrastructure or GPU hardware
  • User needs proprietary closed-source models like GPT-4 or Claude with guaranteed API uptime SLAs
  • User lacks technical expertise to manage model downloads, local server setup, or API integration
  • User requires real-time model updates or automatic model retraining on proprietary data
  • User operates in an air-gapped environment without internet access to download models initially
Commercials

Pricing

Free for home and work use