LM Studio
Run local LLMs, chat with documents, and power apps using a local AI server.
At a Glance
Pricing
Get started with LM Studio at no cost with Run models locally and Document RAG.
Engagement
Available On
About LM Studio
LM Studio is a cross-platform toolkit that lets you download and run open-source LLMs (like Llama, Qwen, DeepSeek, Gemma) directly on your laptop or desktop. You can use it for chatting, retrieving answers from local documents (RAG), or hosting a local LLM server to integrate into your own applications.
It includes a discovery UI for finding Hugging Face models in GGUF or MLX format, a chat interface, document RAG system, and a local server mode for API access. Developers can use the LM Studio SDK (via npm) to build local AI apps without managing model dependencies.
LM Studio is free for both personal and commercial use, with paid plans only needed for enterprise features like private sharing, SSO, and model access control.
Community Discussions
Be the first to start a conversation about LM Studio
Share your experience with LM Studio, ask questions, or help others learn from your insights.
Pricing
Free Plan Available
Get started with LM Studio at no cost with Run models locally and Document RAG.
- Run models locally
- Document RAG
- Local server mode
- Use at home and at work
Enterprise
Enterprise-grade solution with SSO and Private model sharing and dedicated support.
- SSO
- Private model sharing
- Model/MCP access control
- Audit and team management
Teams (coming soon)
Designed for teams with Private artifact sharing and Self-serve collaboration tools and collaboration features.
- Private artifact sharing
- Self-serve collaboration tools
Capabilities
Key Features
- Run open-source LLMs locally on Mac, Windows, or Linux
- Chat interface with support for document RAG
- Host local AI server with REST API
- Discover and download GGUF and MLX models
- Cross-platform support for Llama.cpp and MLX formats
- Free SDK for building local AI apps
- Team and enterprise collaboration via LM Hub (upcoming)