Usage Guide

This guide explains how to interact with Nanobrowser and provides examples of tasks you can perform.

The Side Panel Interface

The side panel is your command center for Nanobrowser. It consists of three main parts:

  1. Header: Provides quick access to start a new chat, view your history, access settings, and join the community.
  2. Message List: Displays the conversation between you and the AI agents, including system status updates, agent plans, and actions taken.
  3. Chat Input: The text area at the bottom where you enter your tasks and commands.

Formulating a Task

To start an automation, simply describe what you want to achieve in natural language. Be as specific as possible for the best results.

Good Example (Specific):

"Find a portable Bluetooth speaker on Amazon with a water-resistant design, under $50. It should have a minimum battery life of 10 hours."

Less Effective Example (Vague):

"Find me a speaker."

Example Tasks

Here are some powerful tasks you can perform with just a sentence:

1. News Summarization

Navigate to a news website and ask Nanobrowser to summarize the latest articles.

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

2. Research on GitHub

Use Nanobrowser to explore repositories based on specific criteria.

Command: Search for the trending Python repositories on GitHub with the most stars.

3. E-commerce Research

Let Nanobrowser find products that match your specific requirements.

Command: Find a portable Bluetooth speaker on Amazon with a water-resistant design, under $50. It should have a minimum battery life of 10 hours.

Follow-up Questions

After a task is complete, Nanobrowser remains in a "follow-up mode." This allows you to ask contextual questions based on the previous actions. For example, after the news summarization task, you could ask:

Follow-up: Which of those headlines is about artificial intelligence?

The agent will use its memory of the previous steps to answer your question.

In-Chat Commands

Nanobrowser supports a few special commands you can type into the chat input:

  • /state: Prints the current browser state to the browser's developer console. Useful for debugging.
  • /nohighlight: Removes any active element highlights from the current page.
  • /replay <historySessionId>: Replays the actions from a previous session. You can find the session ID in the Chat History view.