Airtop: A Browser Automation Tool Using Natural Language Controls

General Introduction

Airtop is an AI-based browser automation tool. It lets users control cloud browsers to perform complex web operations such as logging into a website, crawling data or performing automated tasks through simple natural language commands. It solves the complexity and error-prone nature of traditional scripting, and is especially suited to scenarios that need to deal with multiple authentications (like OAuth or double authentication.) At the heart of Airtop is its cloud browser technology, which can be easily scaled to multiple browser instances to meet the needs of large-scale tasks. It also offers real-time human intervention, allowing users to take over operations when necessary. In short, Airtop makes tedious web automation more intuitive and efficient.

Function List

• natural language control: Users can command the browser in everyday language, such as "logging in to a website" or "grabbing page data".
• Cloud Browser Support: Provides a scalable cloud-based browser for single or large-scale operations.
• Complex authentication processing: Support for OAuth, Double Authentication (2FA) and CAPTCHA resolution for easy access to websites that require a login.
• Data Grabbing and Extraction: Can extract structured data from web pages, such as product information or search results.
• Real-time human intervention: Allows users to take over browser operations at any time with the Live View feature.
• File Manipulation Support: Supports basic functions such as uploading and downloading files, as well as copying and pasting.
• API Integration: SDKs for TypeScript and Python are available for developers to integrate into their projects.

Using Help

Airtop is very easy to use and does not require a complicated installation process, just register for an account and access it through the official website. Below is a detailed description of how to get started and operate its main features.

How to get started

1. Register & Login
   Open https://www.airtop.ai/, click the "Sign Up" button in the upper right corner, and fill in your email and password to complete your registration. After registering and logging in, you will be taken to the Airtop control panel.
2. Creating a Browser Session
   After logging in, click the "Create Session" button, the system will automatically assign a cloud browser for you. You can choose the resolution of your browser (default is human size, e.g. 1920x1080).
3. Using Natural Language Commands
   In the control panel, find the input box and enter the command directly. For example, if you want your browser to open Google, you can type "open Google homepage" and Airtop's AI will parse the command and execute it.

Main function operation flow

1. Natural language control

• procedure::
  Enter a specific command in the input box, e.g. "Search for the latest news from Airtop". The browser automatically opens a search engine, enters the keyword, and displays the results. You can also type "click on the first link on the page" and the AI will simulate human clicking behavior.
• caveat::
  Commands should be as clear and specific as possible, avoiding vague statements. If a command is not executed correctly, you can check the browser status via Live View.

2. Handling of authentication

• procedure::
  If you need to log in to the target website, enter "Open [website address]" first. Then, use the Live View function (click "Generate Live View URL" to generate a link) to log in manually, such as by entering an account password or scanning a QR code. After login, click "Save Session Profile" to save the current session status. After that, you can use the saved session to access the logged in page directly, without having to do it again manually.
• Featured Usage::
  For websites that require double authentication, Live View lets you enter a CAPTCHA in real time. After saving the session, the AI agent can continue with actions such as "Grab my order list".

3. Data capture and extraction

• procedure::
  Let's say you want to capture product information from an e-commerce site, start by typing "Open [e-commerce site] and search for cell phones". After the page loads, type "Extract all product names and prices on the page". airtop will return structured data (e.g. in JSON format), which you can either download directly or fetch via an API.
• Advanced Usage::
  If the page is paginated, you can type "Click on the next page and continue extracting data" and Airtop will automatically turn the page and summarize all the data.

4. Real-time manual intervention

• procedure::
  At any time, if the AI operation encounters a problem, click the "Live View" button to generate a live link. Once the link is open, you can take over as if you were operating a local browser, manually completing complex tasks like solving CAPTCHAs or adjusting page layouts. When you're done, save the session and the AI will continue with subsequent commands.
• Tips for use::
  Live View links can be embedded in your app for easy team collaboration.

5. API integration

• preliminary::
  Download Airtop's SDK (which supports TypeScript and Python) from the official website. Python users can run it for example:

pip install airtop

Then import the library in the code:

from airtop import Airtop

• Example of creating a session::
  Initialize the session with code:

  client = Airtop(api_key="你的API密钥")
  session = client.create_session()
  session.command("打开 https://www.google.com")
  

• Example of extracted data::
  Extract web page data using the API:

  data = session.extract("提取页面上的所有标题")
  print(data)
  

tip

• Getting the API key: After logging in, find the API key in "Settings" in the control panel, copy it and use it for code integration.
• View Document: The official website of the /docs page has a detailed API reference and sample code.
• Saving resources: After you finish using the browser, remember to click "End Session" to close the session to avoid wasting cloud resources.

With these steps, you can easily get started with Airtop, whether it's for simple browsing or complex automation tasks.

application scenario

1. Market Research
   Use Airtop to capture new product information from e-commerce sites, remove advertisement interference, and extract price, model number and other data to easily analyze market trends.
2. Social Media Monitoring
   After logging in to social platforms, posts and comments on specific topics are extracted to help brands understand user feedback.
3. automated office work
   Fill out an online form or download a document with natural language commands to improve daily productivity.

QA

1. What programming languages does Airtop support?
   TypeScript and Python are currently supported, and SDKs for more languages will be added later.
2. Do I need to build my own browser?
   No. Airtop provides a cloud browser that you can simply control via the official website or API.
3. What should I do if I encounter a CAPTCHA?
   After manually entering the verification code with the Live View function and saving the session, the AI can continue the operation.