30
Put AI in your app by tonight. This package provides streaming chat, agents that call your code, live voice, and models that run on the phone itself, all in one package for iOS and macOS.
Any Model, One Line to Change Your Mind: Seamlessly switch between models like Claude, GPT, Gemini, Grok, or Ollama running on your Mac. Your code remains unchanged.
Works Offline: Apple's on-device models answer through the same calls as cloud ones, ensuring functionality without a network connection. No network, no key, no data leaves the phone.
Trustworthy JSON Output: Describe the desired data shape and receive typed Codable values. Invalid output results in a thrown error, preventing app crashes.
Rapid Chat Interface Development: Build a chat screen in an afternoon with ChatSession, which manages messages, streams tokens, and communicates with your existing web app routes.
Real-time Voice Interaction: Engage in live voice conversations over WebSockets. The mic input is processed, speech is outputted, and tools are executed in between. Interruptions are handled gracefully.
Backend Compatibility: The app communicates using the same streaming protocol as your web frontend, allowing you to keep your existing backend infrastructure.
iOS and macOS Native: Leverage Apple Intelligence on-device for free, instant, and private AI responses. Fall back to cloud models with a single line of code when on-device AI is not available.
Add the package, pick a model, and send a message. If Ollama is running on your Mac, you don't even need an API key.
Built with