Hacker News · Feb 14, 2026 · Collected from RSS
Your phone has a GPU more powerful than most 2018 laptops. Right now it sits idle while you pay monthly subscriptions to run AI on someone else's server, sending your conversations, your photos, your voice to companies whose privacy policy you've never read. Off Grid is an open-source app that puts that hardware to work. Text generation, image generation, vision AI, voice transcription — all running on your phone, all offline, nothing ever uploaded. That means you can use AI on a flight with no wifi. In a country with internet censorship. In a hospital where cloud services are a compliance nightmare. Or just because you'd rather not have your journal entries sitting in someone's training data. The tech: llama.cpp for text (15-30 tok/s, any GGUF model), Stable Diffusion for images (5-10s on Snapdragon NPU), Whisper for voice, SmolVLM/Qwen3-VL for vision. Hardware-accelerated on both Android (QNN, OpenCL) and iOS (Core ML, ANE, Metal). MIT licensed. Android APK on GitHub Releases. Build from source for iOS. Comments URL: https://news.ycombinator.com/item?id=47019133 Points: 27 # Comments: 7
Off Grid The Swiss Army Knife of On-Device AI Chat. Generate images. See. Listen. All on your phone. All offline. Zero data leaves your device. Not just another chat app Most "local LLM" apps give you a text chatbot and call it a day. Off Grid is a complete offline AI suite — text generation, image generation, vision AI, voice transcription, and document analysis, all running natively on your phone's hardware. What can it do? Text Generation — Run Qwen 3, Llama 3.2, Gemma 3, Phi-4, and any GGUF model. Streaming responses, thinking mode, 15-30 tok/s on flagship devices. Bring your own .gguf files too. Image Generation — On-device Stable Diffusion with real-time preview. NPU-accelerated on Snapdragon (5-10s per image), Core ML on iOS. 20+ models including Absolute Reality, DreamShaper, Anything V5. Vision AI — Point your camera at anything and ask questions. SmolVLM, Qwen3-VL, Gemma 3n — analyze documents, describe scenes, read receipts. ~7s on flagship devices. Voice Input — On-device Whisper speech-to-text. Hold to record, auto-transcribe. No audio ever leaves your phone. Document Analysis — Attach PDFs, code files, CSVs, and more to your conversations. Native PDF text extraction on both platforms. AI Prompt Enhancement — Simple prompt in, detailed Stable Diffusion prompt out. Your text model automatically enhances image generation prompts. Performance Task Flagship Mid-range Text generation 15-30 tok/s 5-15 tok/s Image gen (NPU) 5-10s — Image gen (CPU) ~15s ~30s Vision inference ~7s ~15s Voice transcription Real-time Real-time Tested on Snapdragon 8 Gen 2/3, Apple A17 Pro. Results vary by model size and quantization. Install Android — Download APK Grab the latest APK from GitHub Releases, install, and start chatting in under 2 minutes. Build from source git clone https://github.com/alichherawalla/offgrid-mobile.git cd OffgridMobile npm install # Android cd android && ./gradlew clean && cd .. npm run android # iOS cd ios && pod install && cd .. npm run ios Requires Node.js 20+, JDK 17 / Android SDK 34 (Android), Xcode 15+ (iOS). See full build guide. Documentation Document Description Architecture & Technical Reference System architecture, design patterns, native modules, performance tuning Codebase Guide Comprehensive code walkthrough Design System Brutalist design philosophy, theme system, tokens Visual Hierarchy Standard Visual hierarchy and layout standards Test Flows End-to-end test flows Test Coverage Report Current test coverage status Test Priority Map Test prioritization guide Test Spec Format Test specification format reference Contributing Contributions welcome! Fork, branch, PR. See development guidelines for code style and the codebase guide for patterns. Acknowledgments Built on the shoulders of giants: llama.cpp | whisper.cpp | llama.rn | whisper.rn | local-dream | ml-stable-diffusion | MNN | Hugging Face Off Grid — Your AI, your device, your data. No cloud. No subscription. No data harvesting. Just AI that works anywhere.