EdgeMind: Offline LLM

Experience sovereign intelligence with EdgeMind — the next-generation, 100% on-device AI assistant engineered by Ndronux Labs. Run state-of-the-art open-source language models directly on your smartphone’s physical processor without sending a single byte of your data to the cloud. EdgeMind gives you lightning-fast responses, complete air-gapped privacy, and zero telemetry. 🔒 100% PRIVATE & AIR-GAPPED BY DESIGN • Zero Cloud Dependency: All inference, memory indexing, and document analysis happen locally on your device's CPU and RAM. • Complete Privacy: Your prompts, chats, questions, and attached files never leave your phone. • Works in Airplane Mode: No Wi-Fi, cellular network, or active internet connection required after downloading your model. • No Tracking: Zero analytics SDKs, zero advertising trackers, and zero third-party telemetry. 🧠 MULTI-MODEL HUB (GGUF NATIVE RUNTIME) • Run leading open-weight LLMs: Meta Llama 3.2, Google Gemma 2, Qwen 2.5, and SmolLM2. • Optimized mobile runtime powered by llama.cpp native FFI with multi-core ARM NEON acceleration. • Easily download and manage quantized GGUF models directly within the built-in Model Hub. 📄 SMART DOCUMENT RAG & PDF CHAT • Attach local PDFs, Markdown notes, text documents, and code files. • Instant Semantic Insights: Extract answers, summarize long documents, and analyze complex information using on-device Retrieval-Augmented Generation (RAG). 🧬 PERSISTENT VECTOR MEMORY • Multi-turn conversations backed by an embedded local vector database. • Automatic session indexing for contextual multi-turn reasoning across your entire conversation. • Seamlessly manage, rename, organize, and export your chat histories. 🧪 PROMPT LAB & CUSTOM PERSONAS • Choose specialized system personas: Software Engineer, Creative Writer, Concise Analyst, or create your own custom instructions. • Fine-tune generation parameters: Temperature, Top-P, Repeat Penalty, and CPU thread allocations to optimize for your phone's hardware. 📊 LIVE HARDWARE & INFERENCE HUD • Monitor real-time performance metrics: Generation speed (tokens/second), time to first token, active RAM footprint, and CPU core utilization. 🛡️ ZERO DATA SHARING Your data belongs exclusively to you. Ndronux Labs does not collect, store, or sell any user information or prompt history. -------------------------------------------------- Developed with pride by Ndronux Labs. Sovereign Intelligence. Total Privacy. Zero Compromises.