Usage

How it works

Once installed and configured, Nanobrowser adds a sidebar to Chrome. You type a natural-language task and the three-agent system (Planner, Navigator, Validator) executes it step by step in your active tab or opens new tabs as needed.

Minimal usage example: extract news headlines

1. Open Chrome and navigate to any page (or a blank tab).

2. Click the Nanobrowser icon to open the sidebar.

3. Type the following task and press Enter:

   Go to TechCrunch and extract the top 10 headlines from the last 24 hours

4. Nanobrowser will:

   • Open https://techcrunch.com in your active tab
   • Scroll and identify article headlines
   • Extract and display the list in the sidebar chat

More task examples

Look for trending Python repositories on GitHub and list the top 5 with their star counts

Find a portable Bluetooth speaker on Amazon under $50 with at least 10 hours of battery life

Go to Hacker News and summarize the top 3 stories

Agent roles

Agent	Role
Planner	Breaks the task into steps; decides what pages to visit
Navigator	Performs browser actions (click, type, scroll, navigate)
Validator	Checks each step outcome; retries or escalates on failure

You can assign different LLM models to each role in Settings. A faster/cheaper model on Validator (e.g., GPT-4o-mini) reduces cost without sacrificing accuracy.

Limitations

• Works only in Chrome/Chromium; no Firefox, Safari, or headless mode.
• Cannot access pages that require a fresh login if your session is not already active.
• Complex multi-tab workflows may time out on slow LLM providers.
• Not suitable for high-volume or automated/scheduled runs; this is an interactive, on-demand agent tool.