Meet the new Locally icon!
With the latest update, we’ve refreshed the app with a new logo that welcomes it into the @lmstudio family.
We’ve also added a side panel for quicker navigation and more Liquid Glass throughout the app.
Locally AI is now available on macOS 💻
Chat privately with local models on your Mac in a beautiful and native app, optimized for Apple Silicon with MLX.
Download on the Mac App Store: apple.co/3IwnTi7
Qwen 3 VL by @Alibaba_Qwen is now available on supported iPhones and iPads.
It’s the most powerful vision-language model in the Qwen family, offering enhanced vision understanding, recognition, and improved OCR capabilities.
Update your app now 📲
The latest update added Gemma 3n E4B for the new iPhone 17 Pro and iPhone Air!
Gemma 3n E4B is a powerful model by @GoogleDeepMind and is designed to be mobile-first.
Locally AI brings Apple’s foundation model to your iPhone, enabling on-device chat, summarization, rewriting, and more. Entirely offline and private.
The model is available on Apple Intelligence-compatible devices.
With the latest iOS, iPadOS, and macOS updates (26.1), the Apple Foundation Model is now available in Locally AI in 6 new languages:
🇩🇰Danish
🇳🇱Dutch
🇳🇴Norwegian
🇵🇹Portuguese
🇸🇪Swedish
🇹🇷Turkish
Download the app: apple.co/3IwnTi7
Introducing Local Voice Mode, chat with any model directly with your voice 🎙️
Local Voice Mode is optimized for natural, on-device, and real-time conversations.
Available today in Preview. Update your app now to get started.
Analyze images with Locally AI on your Mac using the latest Qwen 3 VL models by @Alibaba_Qwen 📷
Running privately on your device and powered by Apple MLX.
Granite 3.3 models by @IBM are now available for iPhone and iPad!
Available in two sizes, a 2B model for iPhones and a 8B model for iPads.
Granite 3.3 delivers efficient performance and reasoning when you need it.
Granite 4.0 by @IBM is now available in the latest update 📣
Granite 4.0 is the next generation models of IBM, featuring a new architecture that greatly reduces memory consumption.
- Micro: a dense 3B model available on all devices.
- Tiny: a mixture of experts 7B (1B active)
You can use Locally AI directly from Spotlight with the "Ask Locally AI" shortcut on macOS, allowing you to quickly get an answer from any model ⚡️
Pro tip: You can assign a quick key to access the shortcut even faster.
Watch it in action 👇
SmolLM 3 made by @huggingface is now available in the app 🤗
- SOTA 3B model
- Dual mode reasoning
- Multilingual (6 languages)
SmolLM 3 outperforms other 3B models while staying competitive with larger 4B models.
Running on-device on iPhone and iPad.
Optimized for Apple MLX.