Locally AI Brings Offline LLMs to iPhone and iPad

Locally AI is an app that lets you run large language models such as Meta Llama, Google Gemma and Qwen directly on an iPhone or iPad — without an internet connection. All data processing happens locally on the device, meaning you can use the app without signing in or connecting to a cloud service.

According to the developer, local processing keeps user data private since no traffic or data is sent to external servers. That makes the app appealing for users concerned about privacy and control over their information.

• Runs popular models (Llama, Gemma, Qwen) on-device
• No internet connection or account required
• Local inference preserves privacy and reduces cloud dependence

Practical considerations: on-device performance will depend on your iPhone/iPad model, available storage, and the specific model you choose to run. Some models may require significant local storage or be available in quantized formats to run efficiently. You may also need to download model files beforehand.

For more details and the original report, see the coverage on iPhone-Ticker or check the app listing on the App Store (search “Locally AI”).

Discussion: Would you prefer an offline local LLM on your phone for privacy, even if it means potentially slower performance or more storage use?