LM Studio
Runs open AI models privately on your Mac
The verdict
Free for home and work use, LM Studio downloads open AI models like Llama, DeepSeek and Qwen and runs them on your Apple silicon Mac in a chat window. Your chats and PDFs never leave the machine, and a local server with OpenAI-compatible endpoints lets other apps on your Mac use that model. It needs Apple silicon and recommends 16 GB of RAM, so Intel Macs and 8 GB machines are the wrong fit.
Get it if
You want to ask AI about private files without sending a single word to someone else's server.
Skip it if
You have an Intel Mac or 8 GB of RAM, or you're happy with ChatGPT in the browser.
Screenshots 1
What it helps you do
- Chat with AI models fully offline
- Ask questions about your own PDFs privately
- Download models like Qwen and DeepSeek in-app
- Run a local AI server for other apps
Plans and prices
LM Studio
FreeFree for home and work use, runs local models on your Mac
Bionic+
$20$20 a month for LM Studio's separate Bionic agent, adds US-hosted cloud models and web search with page extraction
Bionic Pro
$100$100 a month, adds 5 times the Bionic+ usage limits and early access to new features
What changed
Version 0.4.28
Added to The Useful Apps
Written up in The Useful Tech
Pick something else if
ChatGPTPick it if you want 1 assistant you can type to, talk to or show photos on every device.Free to start
ClaudePick this if you write or code for a living and want an AI that reasons through problems with you.Free to start
Draw ThingsPick it if you want to generate and edit AI images on your own device, with control over models and settings.Free to start
The details
- Runs on
- Mac
- Requires
- macOS 12 or later
- Version
- 0.4.28
- Offline
- Works without internet
Developer's websiteSpot a wrong price or fact?