Google LiteRT NeuroPilot: Making Phone NPUs “First-Class Citizens” for On-Device LLMs In the era of pursuing faster, more private AI experiences, running Large Language Models (LLMs) directly on devices is the critical next step. Yet, fitting models with billions of parameters into smartphones and running them smoothly has remained a significant challenge for developers. Recently, the LiteRT NeuroPilot Accelerator stack, launched by Google and MediaTek, aims to turn the NPUs (Neural Processing Units) in MediaTek’s Dimensity series chips into the “preferred target” for on-device LLMs. This is not just another technical update; it seeks to fundamentally change how developers interact …
Imagine telling your phone, “Open Xiaohongshu and find me some weekend travel ideas,” and watching as it silently unlocks, opens the app, taps the search bar, types the query, and scrolls through the results to show you the perfect guide. This scene, straight out of science fiction, is now a tangible reality thanks to the open-source project AutoGLM-Phone-9B. This article will demystify this intelligent agent framework that can “see” your phone screen and “act” on your behalf. We’ll provide a comprehensive, step-by-step guide from zero to deployment, showing you exactly how to bring this automated phone assistant to life. In …
MediaTek NPU × LiteRT: Running LLMs on Phones Without Losing Your Sanity A field-note style walkthrough of the new LiteRT NeuroPilot Accelerator—what it is, why it matters, and how to ship a 1B-parameter model in an Android APK in under 30 min. 0. One-Sentence Take-away You can now compile a Gemma 3 1B model once and run it on millions of MediaTek phones at 1 600 tokens/s prefill—without writing a single line of SoC-specific C++—thanks to the LiteRT NeuroPilot Accelerator. 1. Why On-Device LLMs Keep Getting Stuck 1 cm from the Finish Line Core question: “I already have an INT8 …
Core question of this article: What is GELab-Zero, what problems does it solve in real mobile environments, and why does its design matter for the future of GUI-based mobile agents? This article is a full English rewrite of the selected portions of the original Chinese content. It covers the Background, Capabilities, Application Examples, AndroidDaily Benchmark, and Open Benchmark Results. All content is strictly derived from the provided source file, translated and adapted for a global technical audience. No external facts are added. Table of Contents ☾ Introduction ☾ Why Mobile GUI Agents Matter ☾ What GELab-Zero Provides ☾ Application …
Apple just slipped Model Context Protocol (MCP) support into the App Intents framework in iOS 26.1, iPadOS 26.1 and macOS Tahoe 26.1 dev beta. Translation: ChatGPT, Claude or any MCP-ready model can soon drive your Mac, iPhone and iPad apps—no Shortcuts, no hand-coded REST, no user taps. 1. MCP in One Breath Term Plain-English Analogy Why It Matters Model Context Protocol (MCP) “HTTP for AI tools” One open wire format so every LLM can call any exposed function App Intents iOS’ native “capability outlet” Declare what your app can do; Siri, Spotlight, Shortcuts—and now MCP—can invoke it Apple Intelligence + …
Agent Mode in Android Studio: A Plain-English Guide for Developers How Gemini turns “I want to fix this” into working code—while you keep the final say Table of Contents Why Agent Mode Exists Five-Minute Setup Five Real-Life Examples You Can Copy-Paste What Happens Under the Hood FAQ: The Questions We Hear Every Day When to Use Agent Mode (and When Not To) 1. Why Agent Mode Exists Imagine you are cooking a complicated dish. Old way: wash, chop, season, and stir—every single step by hand. New way: tell a sous-chef, “I want a medium-rare steak with garlic butter,” and watch …
Is There a “Write Once, Run Everywhere” Solution for Android, iOS, and Harmony Next? In today’s mobile landscape, small teams often find themselves spread thin across multiple platforms. Android and iOS have matured ecosystems with robust tooling, but Harmony Next—Huawei’s fledgling multi‑device OS—adds fresh complexity. This case study examines a real-world V2EX post by a two‑person team who needed to ship four cross‑platform apps simultaneously on Android, iOS, and Harmony Next. We’ll translate their experiences and community feedback into an actionable English blog: no extra assumptions, just the raw lessons from the source. 1. Background and Core Challenges A team …
Building Truly Native Android Apps with Swift: The Power of Skip and Swift’s Official Support Developing for iOS and Android using Swift and Skip Breaking Platform Barriers: Swift’s Cross-Platform Evolution Maintaining separate codebases for iOS and Android development creates significant challenges. Swift is breaking down these barriers through Skip tool and Swift’s official Android Workgroup. Developers can now use a single Swift/SwiftUI codebase to build truly native iOS and Android applications. This approach enhances development efficiency while ensuring native performance and user experience on both platforms. Skip: Bridging Swift Code to Android Core Functionality Skip operates through intelligent code transformation: …
Cross-Platform iOS Development Made Easy: A Comprehensive Guide to xtool Introduction: Why Cross-Platform iOS Development Tools Matter Traditional iOS app development requires macOS and Xcode, limiting flexibility and increasing hardware costs. This guide explores xtool, an open-source solution that enables building and deploying iOS apps using SwiftPM on Linux, Windows, and macOS. Section 1: Core Features of xtool 1.1 Cross-Platform Compilation Supports Linux (including WSL), Windows, and macOS Standardized workflows via Swift Package Manager (SwiftPM) Full iOS app pipeline: Compile → Sign → Package 1.2 Developer Services Integration Apple Developer account authentication Automated code signing Device management (install/uninstall/launch apps) 1.3 …