Fullmoon
A local LLM chat app for iOS, iPadOS, macOS and visionOS. Runs on Apple silicon, works offline and saves chat history locally. Licensed under MIT.
Fullmoon is an open source chat app for people who want to run language models on their Apple devices and keep conversations local. It supports iOS, iPadOS, macOS and visionOS, with on-device inference optimized for Apple silicon. You can chat fully offline, and the app saves your chat history locally.
Its model choices include Llama 3.2 1B and 3B, plus DeepSeek-R1-Distill-Qwen-1.5B. Fullmoon supports 4-bit versions of the Llama models and both 4-bit and 8-bit versions of the DeepSeek model. These smaller models make it an option for people looking for local AI chat on an iPhone or iPad as well as a Mac.
Fullmoon also connects local models to Shortcuts. You can use a shortcut to chat with a model or pass its output to other actions, so the app can take part in automation beyond its own chat interface. Appearance settings let you change the theme and fonts, while a custom system prompt lets you adjust how the model responds.
The app uses Swift MLX and Metal for model processing on Apple hardware. It runs inference on your device rather than sending prompts to a cloud model service. Fullmoon's Swift source code is available under the MIT license.
