What is Kuzco?
Kuzco is a Swift package for integrating large language models (LLMs) directly into iOS, macOS, and Mac Catalyst apps. Built on `llama.cpp`, it offers customizable prompts, flexible tuning, and async/await-friendly APIs for on-device AI.
Problem
iOS and macOS developers previously needed to rely on non-native frameworks or build custom solutions to run LLMs locally, resulting in increased development complexity and longer time-to-market.
Solution
A Swift package that enables developers to run LLMs locally on Apple devices with async/await-friendly APIs, allowing integration into apps via native Swift code without external dependencies.
Customers
iOS and macOS developers working on AI-powered apps (e.g., chatbots, content generators) who prioritize on-device processing and Swift-native implementation.
Unique Features
Native Swift implementation optimized for Apple hardware (Metal GPU acceleration), llama.cpp integration, and offline operation without cloud dependencies.
User Comments
Simplifies LLM integration for Apple platforms
Impressive performance on local devices
Essential for privacy-focused AI apps
Reduces development time significantly
Needs expanded model support
Traction
Newly launched (v1.0), open-source package with 1.2k GitHub stars, integrated into multiple private beta apps. Founder actively engaged in Swift developer community (5k+ followers across platforms).
Market Size
The global iOS developer community exceeds 30 million professionals, with mobile AI market projected to reach $26.1 billion by 2028 (MarketsandMarkets).
