Voice Intelligence Navigator

Voice Navigator is an intelligent, voice-guided web navigation assistant built as a Google Chrome extension. Designed to bridge voice interactions with web automation, it allows users to execute complete multi-step goals across web applications through natural conversation. Unlike simple single-action tools, Voice Navigator maintains a persistent, per-tab task state through page reloads, dynamic menu renders, and route transitions. The platform utilizes a dual-planner architecture: a built-in offline local planner optimized for common DOM controls and workflows (such as repository management or file edits), alongside a universal semantic planner that evaluates bounded page snapshots step-by-step using local LLMs (like Ollama's Qwen) or cloud-based AI providers. Voice Navigator processes real-time microphone streams via an offscreen audio pipeline and AssemblyAI's live transcription service. As users speak their goal, the extension inspects the current DOM, visually highlights the next safe interaction, and waits for user confirmation or input before advancing. Built with a strict "human-in-the-loop" security model, Voice Navigator explicitly delegates sensitive steps—such as entering passwords, payment details, authentication codes, or local file selection—to the user, ensuring a safe and controlled guided browsing experience.