timmyy123/LLM-Hub
Local AI Assistant on your phone
What it solves
LLM Hub provides a way to run large language models (LLMs), image generation, and video generation directly on mobile devices (Android and iOS). It eliminates the need for internet connectivity for AI inference, ensuring that user data remains private and conversations never leave the device.
How it works
The app leverages mobile-specific hardware acceleration (CPU, GPU, and NPU) to run models locally. It integrates several specialized SDKs and runtimes, such as GenieX and MediaPipe for Android and Run Anywhere and MediaGenerationKit for iOS. Users can download models directly from HuggingFace or import custom model formats including .gguf, .task, and .litertlm.
Who it’s for
It is designed for users who want private, offline access to AI capabilities on their smartphones, as well as developers interested in on-device AI implementation.
Highlights
- Multimodal Suite: Includes text chat with RAG memory, image generation via Stable Diffusion, video generation via Stable Video Diffusion, and offline translation/transcription.
- Vibe Coder: A tool that allows users to describe an app idea and see it built in real-time with a live HTML/JS preview.
- Privacy-Centric: 100% on-device processing with no data collection, no accounts, and no tracking.
- Broad Model Support: Supports multiple model formats and direct imports from HuggingFace.
- AI Utility Tools: Features a scam detector for phishing analysis, a writing aid for grammar and summarization, and a custom persona creator called creAItor.