Why we built
private AI.
A calmer way to explore AI on iPhone and iPad.
AI can be useful without every question beginning on a remote service. Pebble Mind lets you try supported models on your device, without an account or API key for the local workflow.
A calmer starting point
Pebble Mind is made to feel like an app you can open and use, not an AI infrastructure project. Choose a supported model, let it prepare on your device, and begin with a small task such as summarizing a note, rewriting a message, or brainstorming a checklist.
The local workflow
When you use a supported local model, the conversation runs on the device. Pebble Mind’s privacy description says that chats, prompts, responses, attachments, downloaded models, and preferences are stored locally. No account or API key is required to start that local workflow.
Web search is optional. When you turn it on, a search query goes to the configured search service so the app can retrieve current information and show sources. It is a separate choice from chatting with a local model.
Built for everyday use
Privacy should not require technical vocabulary. Open the app, choose a model that fits, ask a small question, and decide whether the workflow is useful for you. For people who already work with local AI, Pebble Mind also provides a model library, readiness information, local model management, and additional Pro controls.
What we are honest about
Local models use storage, memory, battery, and heat. Larger models may need more of each, and responses can be inaccurate or incomplete. Check important information independently, and choose a model that fits your device.
Pebble Mind is a first version focused on making supported local AI easier to explore on iPhone and iPad, with the privacy boundary visible and web search left to the user.
Try a small question and see whether a local workflow fits your day.
Try Pebble Mind ↗