Overview

banner

Nanobrowser is an open-source AI web automation tool that runs locally in your browser. It serves as a free, privacy-focused alternative to services like OpenAI Operator, featuring a flexible multi-agent system and support for a wide range of Large Language Model (LLM) providers.

Nanobrowser Demo GIF

Nanobrowser's multi-agent system analyzing HuggingFace in real-time, with the Planner intelligently self-correcting when encountering obstacles and dynamically instructing the Navigator to adjust its approach—all running locally in your browser.

Why Nanobrowser?

Nanobrowser offers powerful AI-driven web automation capabilities as a Chrome extension, giving you complete control without the high costs associated with proprietary services.

  • 100% Free & Open Source: No subscription fees or hidden costs. The entire codebase is available for review, ensuring complete transparency. You only pay for what you use with your own API keys.
  • Privacy-Focused: All operations run entirely within your local browser. Your credentials, cookies, and sensitive data never leave your machine.
  • Flexible LLM Options: Connect to your preferred LLM providers, including OpenAI, Anthropic, Gemini, Ollama, Groq, and any OpenAI-compatible custom endpoint. You can assign different models to different agents to balance cost and performance.

Key Features

  • Multi-Agent System: Specialized AI agents (Planner and Navigator) collaborate to accomplish complex web automation workflows.
  • Interactive Side Panel: An intuitive chat interface provides real-time status updates and allows for interactive control over tasks.
  • Task Automation: Seamlessly automate repetitive tasks across various websites with simple, natural language commands.
  • Follow-up Questions: Engage in a continuous dialogue with the agent, asking contextual follow-up questions about completed tasks.
  • Conversation History: Easily access, manage, and replay your interaction history.
  • Multi-LLM Support: Connect to a variety of LLM providers and assign specific models to different agents based on your needs.

Ready to get started? Head over to the Installation guide.