skills.rest
mouchegmouradian
mouchegmouradianCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

zetic-mlange

Integrate ZETIC Melange on-device AI inference into Android and iOS apps.

2|Updated Mar 2, 2026
Search Tags:#ios#android#inference#on-device#zetic#melange#mlange
One-click install
npx skills add https://github.com/mouchegmouradian/claude-code-skills --skill zetic-mlange
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: zetic-mlange
Source: https://github.com/mouchegmouradian/claude-code-skills/tree/main/skills/zetic-mlange
Command: npx skills add https://github.com/mouchegmouradian/claude-code-skills --skill zetic-mlange

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Enables seamless integration of the ZETIC Melange on-device AI SDK into Android and iOS apps, giving Claude Code deep context on how to run on-device models with correct concurrency, lifecycle management, and hardware targets.

Core Features & Use Cases

  • General model inference with ZeticMLangeModel on Android (Kotlin/Java) and iOS (Swift)
  • LLM inference with streaming tokens using ZeticMLangeLLMModel
  • HuggingFace model loading via ZeticMLangeHFModel
  • Real-time, thread-safe mobile inference following zero-allocation patterns and lifecycle cleanup

Quick Start

Initialize Melange in your mobile app and run a simple on-device inference following the threading rules.

Frequently Asked Questions about zetic-mlange

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run on-device model inference in Android and iOS without blocking the main thread?▼

On-device model inference requires blocking JNI or Swift calls moved off the main thread, using mutexes or actors for serialized access, and proper lifecycle cleanup to prevent memory leaks during hardware backend handling.

Can I load HuggingFace models for on-device inference in mobile apps?▼

Yes, HuggingFace model loading is supported on both Android and iOS via ZeticMLangeHFModel, enabling direct integration of HuggingFace models into your mobile app for on-device inference with proper threading.

How do I stream LLM tokens on-device for Android and iOS?▼

LLM streaming inference is handled by ZeticMLangeLLMModel, providing real-time token streaming on Android and iOS while maintaining zero-allocation hot paths and thread-safe execution.

What lifecycle management rules are required for on-device mobile AI inference?▼

On-device mobile AI inference requires proper lifecycle cleanup to manage hardware backends, serialized access via mutexes or actors, and zero-allocation hot paths to ensure reliable model execution without memory leaks.

Does Zetic Melange support both Kotlin and Swift for on-device model inference?▼

Yes, Zetic Melange supports on-device model inference across Android using Kotlin or Java and iOS using Swift, providing platform-specific setup guidelines for hardware backend handling and thread-safe execution.

Related Skills

Explore More Software Engineering Skills→
vertivolatamvertivolatam
1

Native Integration (Swift/Kotlin)

Integrate Flutter apps with native iOS Swift/UIKit and Android Kotlin/SDK features.

Official
Advanced
dankoflydankofly

full-stack

Coordinate backend and frontend changes for end-to-end features.

Community
Advanced
ai-enhanced-engineerai-enhanced-engineer
3

react-native-ml-inference

Manage TensorFlow.js tensor lifecycle and normalize Teachable Machine inputs in React Native.

Official
Intermediate
plurigridplurigrid
60

borkdude

Guide ClojureScript runtime selection across scripting, browser, and Node contexts.

Official
Advanced