Locally AI: run models on your iPhone or iPad

Locally AI is an AI chat app for iPhone and iPad that runs models on the device. Product information describes offline use, text chat, image analysis and text generation. It may interest people who want to try AI without a connection or prefer processing content locally.

Models and use

The listed model families include Llama, Gemma and Qwen, with examples such as Llama 3.2, Google Gemma 2 and 3, and Qwen 3. The exact catalog, download size, device compatibility and speed may vary by app version and hardware. Check the current in-app model list before downloading a large model.

The app is described as supporting chat, image analysis and text generation. Its description also says that no login is needed and data stays on the device. Verify privacy statements against the current policy, permissions and system settings. Local storage alone does not protect data if a device is lost or shared.

To get started, install the app, check model and device requirements, prepare a model, then try chat, text generation and image analysis. Model files may use substantial storage. Available information does not specify pricing, minimum OS versions, a complete model list, exact device requirements, backup behavior or detailed privacy terms. Offline inference does not necessarily mean downloads, updates or every feature work without a network.

FAQ

Models are described as running locally, but downloads or updates may need internet. Llama, Gemma and Qwen are listed; check the app for current options. The description says data remains on-device, but verify current privacy information before adding sensitive content.