AI LLM Gallery: Local AI Chat

AI LLM Gallery is a private AI chat app and model gallery that runs language models locally on your phone. Download an open model such as Gemma or Qwen once, then chat, write and ask questions about your documents offline, even in airplane mode. No account, no cloud AI, no ads. WHY AI LLM GALLERY • Offline AI chat: once a model is downloaded, every reply is generated on your phone's own hardware. • Private by default: your chats, documents and prompts stay on your device. There is no account, no backend server, no analytics and no ads. • Internet only when you ask for it: to download models, and for the optional web Research tool, which is off until you turn it on. CORE CAPABILITIES • Private AI Chat: Natural, streaming conversations with complete local history, custom system prompts, temperature and sampling controls, code syntax highlighting, and transparent reasoning display. • Model Gallery & Hub: Curated catalog of open models (Gemma, Qwen and more) with "fits my device" badges computed from your phone's available RAM. Supports GGUF and LiteRT-LM formats with one-tap downloads. • Knowledge Documents: Import PDFs, text files or notes for private, on-device Q&A with citations. • Writing Studio: Distraction-free writing assistant for drafting, rewriting tone and summarizing. • Research (optional): Look something up on Wikipedia, or with your own Brave Search key, and get a cited answer. When you use it, your search question is sent to that service. • Prompt Lab & Benchmarks: Test prompts across templates, run hardware speed benchmarks (tokens/sec), and inspect throughput. BUILT FOR MOBILE HARDWARE • Runs on CPU or GPU, with an NPU option on supported phones. • Resumable, verified model downloads with Wi-Fi-only and mobile data warnings. • Storage manager to inspect and remove models at any time. • Material 3 design with light and dark themes. RESPONSIBLE AI Responses are generated by on-device language models and may be inaccurate. An AI disclaimer is shown on first launch, and every message has a report action.

Initial release of AI LLM Gallery - Private On-Device AI Workspace. - On-device local LLM chat and reasoning - Curated model hub with one-tap downloads - Offline document Q&A and writing assistant - Zero latency private inference