br-AI-n

A fully private, offline AI assistant running local LLMs in the browser using WebAssembly and WebGPU.

The ultimate privacy-first AI companion built directly into your Chrome Side Panel. br-AI-n is a powerful extension that allows you to chat with Large Language Models (LLMs) and Vision Models privately. By leveraging state-of-the-art WebAssembly and WebGPU hardware acceleration, your AI assistant runs directly on your local machine without the need for API keys, monthly subscriptions, or hidden telemetry. 🛡️ PRIVACY BY DEFAULT & LOCAL-FIRST The extension does not collect, store, or transmit your chat history, screenshots, or usage data to developer-owned servers. • In-Browser Models: Your sensitive documents, code snippets, images, and chats are processed entirely inside your browser's local sandbox. • Local Servers: If you choose to connect to a Local Native Server, your data is sent only to the local endpoint you specify. • (The only time outbound requests occur is if you explicitly toggle on Web Tools to search the internet or when downloading open model weights into your browser's local cache). 🚀 HIGH-PERFORMANCE NATIVE EXECUTION • WebGPU Accelerated: Harnesses the power of your graphics card for blazing-fast inference speeds. • CPU Fallback: Older hardware? No problem. The extension falls back to multi-threaded CPU processing. • Instant Caching: Download models once. They are safely cached in your browser's IndexedDB for instant, offline loading in the future. • Open-Weights Attribution: Full transparency with built-in creator and license info (Apache 2.0, MIT, Gemma) linking directly to official Hugging Face model cards. 👁️ OFFLINE MULTIMODAL VISION & TAB ANALYSIS Analyze visual content with complete privacy using compact, local vision models: • Instant Tab Capture: Click "Capture Tab" to take a screenshot of your active webpage and ask the AI to explain diagrams, charts, infographics, or UI layouts. • Image Inspection: Attach local image files or right-click images on any webpage to inspect and chat about visual content completely offline. 🔌 BRING YOUR OWN LOCAL SERVER Already running a powerful local AI setup? br-AI-n seamlessly connects to Local Native Servers (like Ollama, LM Studio, or Llama.cpp) running on your local network. Offload heavy processing to your dedicated rig while enjoying the sleek, persistent Chrome interface! 🎭 FULLY CUSTOMIZABLE PERSONAS Tailor the AI to fit your exact needs! br-AI-n comes with built-in presets (Assistant, Code Ninja, Grammar Checker, Summarizer, Translator). You can also effortlessly create, edit, and save custom Personas with tailored system prompts to dictate exactly how the AI behaves and formats its answers. 📝 FLOATING INLINE WIDGET Highlight any text on a webpage and br-AI-n's sleek inline pop-up widget will appear! • Instantly apply your favorite personas (like Summarize, Explain Code, or Grammar Check) directly on the page. • Seamless Model Sharing: The in-page widget and Side Panel seamlessly share your active local model without conflicting or reloading. • The inline widget remembers context: type follow-up questions in the built-in chat input to seamlessly continue the conversation right where you are reading. 🧠 AUTONOMOUS AI AGENT TOOLS Transform your local LLM into a powerful research agent! When you need real-time facts or external information, turn on Web Tools to fetch web pages and search the internet. (Note: Web Tools are disabled by default for 100% offline privacy. You can easily toggle them on-demand directly from the prompt toolbar when you need live search capabilities). 📄 SMART WEB CONTEXT INTEGRATION Reading an incredibly long article? Click the "Attach Page Context" button, and br-AI-n will instantly scrape the active tab and feed the text directly into your local LLM's context window. Perfect for instant summarization, extracting key points, or asking Q&A about the page you are currently viewing. ✨ DESIGNED FOR MODERN WORKFLOWS • Reasoning Model Support: Natively parses and renders "Thought Processes" from advanced reasoning models (like DeepSeek-R1 derivatives), elegantly streaming their thoughts into interactive drop-down boxes. • Persistent Sidebar: Designed to live in your Chrome Side Panel so your AI is always one click away, no matter what tab you are on. Take back control of your data. Experience the future of decentralized AI right inside your browser with br-AI-n.