Vicaro

Vicaro is a self-hosted browser agent for Chrome that reads your open tabs, understands each page, and completes tasks for you. Key features: • Reads open Chrome tabs and understands page content • Clicks, types, and fills forms in accounts you are already signed into • Runs on your Mac through a local helper that handles the task loop • Connects to Ollama, LM Studio, or any OpenAI-compatible endpoint you choose • Pauses for your approval before any irreversible action • Works from a side panel inside the browser you already use • Currently in beta and available for macOS only. Vicaro is designed to take over repetitive browser work while you stay in control. Instead of copying information from one page into another, you can ask it to research, draft, fill out forms, or complete multi-step tasks directly in the tabs you have open. Because it uses your existing sign-ins and browsing context, it does not need separate credentials or integrations to start working. Privacy is built into the architecture. There is no Vicaro server in the loop: the helper, history, and memories stay on your machine, and page content is sent only to the model endpoint you configure. You decide which model runs the work, whether that is a local server or a hosted OpenAI-compatible service. Because the beta can click, type, submit forms, make purchases, delete data, and change account settings, reviewing each action before approval is strongly recommended. Vicaro suits professionals, researchers, and privacy-conscious users who spend much of their day in Chrome and want an assistant that keeps data on their own hardware. Common use cases include filling out administrative forms, gathering information across several tabs, handling routine account tasks, and drafting content directly in web apps. Requires macOS 13.5 or later and Chrome 116 or later. Free during the beta period with no Vicaro account required.