sonuai.dev logo
SonuAI.dev
AI

Browser-Use Free AI Agent: Now AI Can control your Web Browser

Browser-Use Free AI Agent: Now AI Can control your Web Browser
0 views
5 min read
#AI
Table Of Content

In this article, I’ll walk you through an incredible open-source tool called Browser Use, which allows AI agents to control and automate tasks in your browser. This tool is not only powerful but also easy to use, making it a great option for anyone looking to automate web-based tasks.

What is Browser Use?

Browser Use is an AI-powered browser automation framework that connects AI agents to your browser. It provides a simple yet powerful interface for automating tasks like web scraping, interacting with websites, and using large language models.

Key Features of Browser Use

  • Vision + HTML Extraction: Combines visual and HTML data for better automation.
  • Multi-Tab Management: Handles multiple browser tabs efficiently.
  • Element Tracking: Tracks and interacts with specific elements on a webpage.
  • Support for Large Language Models: Works seamlessly with models like OpenAI and Anthropic.

Browser-Use AI Agent Overview

DetailInformation
NameBrowser Use
TypeAI-powered browser automation framework
Accuracy89% (web agency benchmark)
GitHub Repobrowser-use/browser-use
Supported ProvidersOpenAI, Anthropic
Open-SourceYes
Key FeaturesVision + HTML extraction, multi-tab management, element tracking, LLM support

How to Install Browser Use?

Prerequisites

Before installing Browser Use, ensure you have the following:

  1. UV: For setting up a virtual environment.
  2. Python 3.11 or above: Required for running Browser Use.
  3. Playwright: A browser automation library.

Installation Steps

  1. Set Up the Virtual Environment:

    • Open your command prompt and paste the following command to create a virtual environment: 
      python -m venv browseruse_env
    • Activate the environment by running: 
      source browseruse_env/bin/activate  # For macOS/Linux  
browseruse_env\Scripts\activate     # For Windows

  2. Install Dependencies:

    • Install Browser Use by running: 
      pip install browser-use
    • Install Playwright: 
      playwright install

  3. Clone the Repository (Optional):

    • If you want to use pre-built templates, clone the Browser Use repository: 
      git clone https://github.com/browseruse/repo.git
    • Navigate to the examples folder to explore and use the templates.

  4. Set Up API Keys:

    • Rename the .env.example file to .env.
    • Paste your API key for your preferred provider (OpenAI or Anthropic).

Demo Examples of Browser Use in Action

1. Automating Job Applications

In one demo, Browser Use automates the process of applying for jobs. Here’s how it works:

  • The AI agent reads a CV and extracts relevant keywords like Python and TensorFlow.
  • It searches for machine learning jobs on platforms like LinkedIn and Indeed.
  • The agent scrapes job postings, saves the results in a structured file, and opens application pages in new browser tabs.
  • It can even autofill details for you!

2. Finding Flights on Kayak

Another demo shows Browser Use finding flights on Kayak:

  • The agent inputs the departure (Zurich) and destination (Beijing) along with the travel date.
  • It scrapes flight options, including prices and schedules.
  • The data is saved for review or refinement, as seen on the right-hand side of the demo.

3. Finding Models with a Specific License

In this demo, the agent is tasked with finding the five most-liked models with a specific license on Hugging Face:

  • It starts by handling the Google search cookies pop-up.
  • The agent navigates to Hugging Face, applies the correct filters, and extracts the models.
  • It calls a custom function (save_models) to save the model details (title, URL, likes) in a structured format.

Why Browser Use Stands Out

Accuracy and Speed

Browser Use has been tested in a web agency accuracy benchmark test, where it outperformed other tools like Web Voyager, Computer Use Agent E, and Ronner H. It achieved an impressive 89% accuracy, making it one of the most reliable AI agents for web-based tasks.

Open-Source and Easy to Install

Browser Use is fully open-source, meaning you can access and modify it freely. Installing it is straightforward, and I’ll guide you through the process step by step.

Using Browser Use: Examples

Example 1: Multi-Tab Handling

  • Task: Open three tabs with searches for Elon Musk, Trump, and Steve Jobs.
  • Result: The agent opens all three tabs, navigates back to the first tab (Elon Musk), and stops.
  • Task: Search for a laptop on Amazon, sort by best rating, and retrieve the price of the first result.
  • Result: The agent navigates to Amazon, performs the search, and returns the price of the top-rated laptop.

Example 3: Multiple Agents Working Together

  • Task: Open two tabs with Wikipedia and deploy another agent to analyze the content.
  • Result: The agents work autonomously, handling multiple tasks simultaneously.

Creating Your Own AI Agent

Creating a custom AI agent with Browser Use is simple:

  1. Choose Your Provider: Select between OpenAI or Anthropic.
  2. Define Your Task: Replace the task in the template with your desired action (e.g., searching flights, buying something online, or web scraping).
  3. Run the Agent: Execute the script and watch your agent automate the task.

Final Thoughts

Browser Use is an incredibly powerful and flexible tool for automating web-based tasks. It can scraping data, filling out forms, or managing multiple tabs, this open-source AI agent can handle it all with impressive accuracy and speed.

If you’re interested in trying it out, I’ve included all the necessary details related to the Browser-use ai agent.

That’s it for today’s guide! I hope you found this article helpful and are excited to explore the possibilities with Browser Use.

Related Posts