Nanobrowser — AI Web Automation Chrome Extension

What it is

Nanobrowser is an open-source Chrome extension that brings multi-agent AI to web automation. Instead of writing scripts or recording macros, you describe what you want done in natural language and the AI agents navigate, click, fill forms, and extract data from web pages autonomously. It serves as a free alternative to commercial browser automation tools.

Users who need to automate repetitive web tasks -- data extraction, form filling, testing, or multi-step workflows -- without writing code benefit most. The multi-agent architecture means different agents handle navigation, extraction, and validation in parallel.

How it saves time or tokens

Nanobrowser eliminates the need to write and maintain browser automation scripts. Traditional tools like Selenium or Playwright require code for every interaction. Nanobrowser translates a single natural language instruction into a sequence of browser actions. The multi-agent design handles dynamic pages, pop-ups, and authentication flows that break simple scripting approaches. Updates to website layouts do not require script maintenance because the AI adapts to visual changes.

How to use

1. Install Nanobrowser from the Chrome Web Store or build from source:

git clone https://github.com/nicepkg/nanobrowser.git
cd nanobrowser
npm install && npm run build

1. Load the extension in Chrome via chrome://extensions with Developer Mode enabled.

1. Open the Nanobrowser sidebar and describe your task:

Go to amazon.com, search for 'mechanical keyboard', and extract the top 5 product names and prices

The agents navigate the page, perform the search, and return structured results.

Example

# Task: Fill out a multi-step form
Nanobrowser prompt:
'Go to forms.example.com/application,
 fill in name as John Smith,
 email as john@example.com,
 click Next,
 select Experience Level: Senior,
 click Submit'

# Task: Monitor a page for changes
'Check news.ycombinator.com every 5 minutes,
 alert me if a post about AI regulation reaches the front page'

# Task: Extract data into a table
'Go to github.com/trending, extract repo name, stars, and language for all repos on the page'

Related on TokRepo

• Browser Automation Tools -- AI-powered browser automation on TokRepo
• Web Scraping Tools -- Data extraction and scraping utilities

Common pitfalls

• Nanobrowser requires an LLM API key (OpenAI, Anthropic, etc.) for the AI agents. Each automation task consumes tokens proportional to the complexity of the page and task.
• CAPTCHAs and anti-bot measures can block automated interactions. Nanobrowser works best on internal tools or sites without aggressive bot detection.
• Complex multi-step workflows may require breaking the task into smaller instructions. A single prompt that tries to do too much can confuse the agent planning.