LM Studio

• Local model hosting: Run models like Llama, Qwen, Gemma, DeepSeek, gpt-oss, and more directly on your machine.
• Built-in catalog: Discover and download models directly in the LM Studio app.
• OpenAI-compatible API: Local API/server supporting chat, completions, embeddings, and image inputs (where supported).
• Offline RAG: Attach .docx, .pdf, or .txt files to chats for fully offline retrieval-augmented generation.
• Presets & Hub: Bundle prompts and parameters into reusable presets; share via LM Studio Hub.
• MCP plugins & SDKs: Python/JS SDKs and plugin support for tool use and workflows.
• Headless mode: Run as a background service for agents and automation.
• Cross-platform: macOS (Apple Silicon), Windows (x64/ARM), Linux (x64) with llama.cpp/MLX optimization.

Individuals and teams who want on-device AI with privacy by default, and enterprise customers seeking secure deployment, SSO, and admin controls.

A privacy-first, cross-platform local AI studio. LM Studio blends a friendly chat UI, offline RAG, OpenAI-compatible API, and preset sharing into a single package—ideal for developers and teams that want local control without sacrificing usability.

• Private and offline by design—no internet required for core functions.
• OpenAI-compatible API reduces integration effort.
• Cross-platform with hardware-optimized runtimes.
• Presets and Hub streamline reuse and sharing of prompt configurations.
• SDKs and headless mode enable broader automation and local agent use cases.

• Performance depends heavily on local hardware specs.
• Teams plan not yet launched; advanced collaboration may require the Enterprise tier.
• REST API is still in beta and may change.

‍

Community (Free)

Desktop app for personal and work use; public Hub sharing.

Teams (Launching soon)

Private Hub orgs; team collaboration features.

Enterprise (Custom)

SSO; model gating; admin controls; custom deployment; priority support.

‍