UnPlugged LM

Your AI. Your data. Your device. No cloud. No subscriptions. No compromise. UnPlugged LM is a fully private, offline AI assistant powered by llama.cpp. Every inference runs directly on your Android device — no internet required, no data ever leaves your phone. ────────────────────────────── FREE FEATURES ────────────────────────────── PRIVATE AI CHAT Stream responses from open-weight large language models running entirely on your CPU. Full chat history saved locally. New Chat button to start fresh any time. DOCUMENTS & RAG (Retrieval-Augmented Generation) Import PDFs and Word documents. UnPlugged LM reads, chunks, and embeds them on-device so you can ask questions grounded in your own files — contracts, medical notes, research, receipts, anything. PROJECTS Organise documents and conversations into separate workspaces. Each project maintains its own RAG context so topics stay isolated. MODEL MANAGER Download and manage GGUF models directly inside the app — Phi-3 Mini, Llama 3, Gemma 2, Mistral, and more. Models live in private app storage. Switch between them any time. Loading indicator shows progress as a model is loaded into memory. ────────────────────────────── PRO — $9.99 ONE-TIME PURCHASE ────────────────────────────── LOCAL AI SERVER Turn your Android phone into a local AI server on your Wi-Fi network. Exposes an OpenAI-compatible API (port 11435) so any app or device on the same network can use your phone's AI — no cloud, zero cost. Includes: • Built-in browser chat UI — open from any browser on your network • Optional access code to gate who can connect • Foreground service keeps the server alive when the screen is off • Browser chat sessions saved to your phone's chat history More Pro features rolling out soon: Voice I/O, Device Actions, Scheduled Tasks, Morning Briefing, Smart Notifications, Unlimited Projects. ────────────────────────────── PRIVACY — OUR CORE PROMISE ────────────────────────────── ✓ Zero telemetry — no analytics, no crash reporters, no trackers ✓ No account, no sign-up, no email required — ever ✓ All AI inference runs on-device via llama.cpp (C++ native) ✓ Documents stay in your app's private sandboxed storage ✓ The local Wi-Fi server never reaches the internet ✓ Open-weight models you download are yours to keep Built for lawyers, doctors, researchers, journalists, and anyone who handles sensitive data and can't afford a privacy leak. My Device. My Rule. Developed by SmartDataIQ — unpluggedlm.com