Waffy - AI Copilot for Web

Automate your browser with AI. Browse, click and complete multi-step workflows with models you like. Free and open source.

Waffy is a free, open-source browser extension that automates web tasks using AI. Runs entirely in your browser with no backend or account required. Tell it what you want, and it will browse pages, click buttons, fill forms, scroll, navigate and complete multi-step workflows for you. You can also use on-device browser models (like Gemini Nano in Chrome) for free. Use it to automate repetitive tasks like filling forms, researching across sites or extracting information from a page. It reads the screen, plans the steps and carries them out while you can continue with other work. You can also chat with it, ask questions about the page, get summaries and even upload files or images for context. Beyond chat and automation, Waffy includes Search and Research modes. Search looks up live answers from the web. Research digs into the current page and the web when needed — to summarize, extract and synthesize deeper findings. Open Waffy from the toolbar or press Ctrl+I (Cmd+I on Mac) to use the side panel alongside your current tab. Waffy is standalone: all extension code runs locally, and your API keys, settings, and conversation history stay in your browser. Connect your own keys from providers like OpenAI, Anthropic, Google, etc. Requests go directly from your browser to the provider you choose. Waffy does not proxy, host or log your data. Waffy also supports browser built-in models. Download the model once and it runs locally on your device — no API key, and nothing leaves your browser. Use it for simple tasks and mix it with cloud models for heavier reasoning. No coding or scripts needed. Just describe what you want done. You stay in control because everything runs on the sites you are already using. Choose different models for planning, execution, and validation to balance speed, cost, and capability. Contribute and star on GitHub: https://github.com/upzare/Waffy