LM Studio is a desktop app for running large language models (LLMs) locally on your own computer.
features
- Run Local LLMs Offline — Execute open-weight models entirely on your own device without cloud dependency.
- Cross-Platform Support — Available on Windows, macOS, and Linux.
- Model Discovery & Download — Browse and download GGUF/MLX models directly from Hugging Face.
- Built-in Chat Interface — Interact with local models through a clean ChatGPT-like UI.
- OpenAI-Compatible API Server — Serve local models via OpenAI-style REST endpoints for app integration.
- Document Chat (RAG) — Chat with local documents privately and offline.
- Model Management — Organize models, prompts, presets, and configurations in one place.
- Developer Tooling — Includes CLI, Python SDK, TypeScript SDK, and REST
- Headless Mode — Run LM Studio without GUI for servers or CI environments.
- MCP Integration — Connect MCP servers and external AI workflows.
- Privacy-First Design — Prompts and conversations stay on the local machine.
- Hardware-Aware Optimization — GPU offloading and memory controls for better performance on limited hardware.
- Multi-Model Support — Supports Llama, Qwen, Mistral, Gemma, Phi, DeepSeek, and more.
- Local Network Serving — Expose local models across devices on the same network.
