ChatGPT web scraping (with a bot)

Some people fear AI taking over 😱, while others use it to automate their jobs 🕺. How do you feel about it? If you're looking to gain skills to automate and build bots with powerful AI capabilities, you've come to the right place. We'll show you how to create a bot that utilizes ChatGPT AI to scrape data websites.

If you want to dive straight in, why not start with the ChatGPT web scraper template. (opens new window)

# Why would you want to scrape websites with a ChatGPT bot?

There are lots of good no-code web scrapers like and Octotparse. They are great at scraping content from websites if they are always in the same format. We call that structured data. But 💁🏼 sometimes you want to scrape data when it's unstructured, when it doesn’t appear in a consistent format or place and we need the help of an AI to find it.

# What scraping use case would you recommend using a ChatGPT bot for?

The primary use case is if you want to scrape many different websites with a single bot, looking for a defined data set. The data would appear unstructured as each website is different, and this is where an 🧠 AI scraper bot would be a useful tool. It would allow you to scrape thousands of pages for phone numbers or emails, for example. Great for lead building.

# What exactly is a ChatGPT bot web scraper?

🧠 ➕ 🤖 A ChatGPT web scraper bot combines browser automation technology and AI ChatGPT tech. It's like a robot, but for the browser.

# How do you make a ChatGPT scraper?

With Axiom's no-code bot builder 🛠️ , you can create bots for web scraping without having to code. Use a simple point and click interface to create as many ChatGPT scraper bots as you want! If you follow the steps below, you'll have a working bot doing your web scraping in no time.

# How does a ChatGPT scraper work?

The scraper works by looping through a list of URLs from a Google Sheet, visiting each page, and extracting the data you need, using its AI capabilities. The bot then writes the data to another Google Sheet, leaving you free to sit back and enjoy a cup of coffee!

# Do I need my own Chat GPT API key?

You probably already have one, but they are simple to get hold of if you don't. Just follow 🔐 OpenAI's instructions. (opens new window)

The short answer is that there’s no law or rule banning web scraping, but that doesn’t mean that you can scrape everything. Web scraping is completely legal if you scrape data that is publicly available online. (opens new window)

Certain types of data however are protected by international regulations, so be mindful of what you can and can’t scrape when it comes to personal data, intellectual property, or confidential data.

# Let’s learn how to build a ChatGPT scraper bot with

If you want to short cut the process why not try the template or read on to learn more!

# 1. Set up your Google Sheet

🏁 Create a new Google Sheet. You can do this in your Chrome browser by entering '' into the address bar. Don’t forget to name your sheet something like 'Websites to scrape’. Add two tabs; one named ‘URLs’ and another named ‘Data’.

# 2. Start from blank getting started making a bot

To build your bot from scratch, click on 'Add first step’. This will open the step selector and you can start adding steps to your bot.

# 3. Add your first step: ‘Read data from Google Sheet’

Use the Step Finder to search for ‘Read data from Google Sheet’ and click on it. The step will be added to Axiom for you to configure. configuring read data from Google Sheet step

In the field called 'Spreadsheet URL', you can search for the Google Sheet you created by name. Once found, click on it to select.

For 'Sheet name' click on the drop-down and select the correct tab.

In the 'First cell' field, toggle the switch and enter 'A1’. This setting tells the bot where to start reading data.

In the 'Last cell' field, click the toggle switch and enter 'A10'. You have limited the bot to read ten rows. This is fine for now, you can increase the amount later once you’ve tested it.

If you want to learn more about Google Sheet steps, watch these videos (opens new window).

# 4. Add the ‘Loop through data’ step find loop step in finder

Next, add a new step by entering ‘Loop through data’ into the Step Finder, and adding it. This step will allow your bot to loop through the rows of data stored in the Google Sheet. Make sure to add the steps you want to repeat inside the loop. configure loop step

# 5. Add the ‘Go to url’ step inside the loop

Add the 'Go to page' step, in the 'URL’ field, select ‘Insert data’ and select the data from your Google Sheet, so that the bot knows which data to loop through. pass data ionto a go to url step

# 6. Add the ‘Get data from bot's current page’ step configure the scraper

Because we want to scrape many different webpages, we need to use a custom selector to scrape a common element found on every page. This common element will be ‘body’. Select it by clicking ‘Select’ and in column A, selecting ‘Custom selector’ and typing in the word ‘body’. Then click on ‘Confirm selector’ and then ‘Complete’. adding a custom selector

Set the ‘Max results’ setting to ‘1’.

# 7. Add the ‘Extract data with ChatGPT’ step

You will need to enter your API key for this step to work, so make sure you have it to hand. adding the ChatGPT step

Then, in the Data field, set it to ‘Scrape data’.

Next, in the ‘Extract values’ field, tell the AI what data you want to extract, by entering a comma-separated list of items, for example email, address, dates, URL etc. You may want to experiment a little here.

# 8. Write data to a Google Sheet

Now we’re going to write the data that ChatGPT AI extracted, to a Google Sheet. write data to a google sheet step

Click 'Add a new step'. In the 'Spreadsheet URL' field, search for the sheet you created at the beginning of this tutorial.

In the 'Sheet name' field, select 'Data'.

In the 'Data' field, select '[chat-gpt]'.

Finally, in the ‘Write options’ field, click the ‘Add to existing Data’ option.

# 9. Within the loop step, add ‘Delete rows from a Google Sheet’ step delete rows from google sheet step

Add the step 'Delete rows from a Google Sheet' to delete the URL just scraped. This is to prevent the same row from being scraped repeatedly. Using the Step Finder, search for and select ‘Delete rows from a Google Sheet’.

In 'Spreadsheet URL' click to add the spreadsheet.

Then, in ‘First row to delete' enter the number 1 and repeat this in 'Last row to delete', so that both are set to '1'. We only want to delete one row per loop.

Set the ‘Sheet name’ to the original spreadsheet set up.

# 10. Ready to test

👩‍🔬 Now you’re ready to test with the desktop app, by clicking ‘Run’.

When you’re done testing and you want to increase the size of your loop, edit the ‘Last cell’ field mentioned in step 3, and run again.

# 11. Running the bot

🚀🚀🚀🚀🚀 You can run this bot in the cloud and the desktop app. If you want to learn more about scheduling, click here..

# Design pattern for your bot

Your assembled ChatGPT bot should resemble this image. design pattern for a ChatGPT web scraper

# Issues you may encounter with your ChatGPT web scraper

  1. The bot only scrapes 10 rows, In the 'Read data from Google Sheets' change the 'last cell' value
  2. ChatGPT returns no data, try adjusting your extract values
  3. No data or incorrect data being written to Google Sheets. Check the ChatGPT data has been inserted into the 'Write data to a Google Sheet' step

Don't forget we offer excellent customer support. If you need help, get in touch. (opens new window)

# Conclusion

Congratulations 🥳👨‍🎓🥳, you've learned how to make and use a ChatGPT AI bot to extract data. With this newly acquired skill, you know how to scrape data, loop through actions, and extract data to a Google Sheet. The sky's the limit with your new AI super powers. 🦸

# What else can I automate with ChatGPT and Axiom?

If you're excited, here are some ideas for other bots: extract data from your email app (such as Gmail), generate content, and send direct messages on Instagram. We have steps to extract data with AI and to generate text with AI. Take a look at our templates.


    Install the Chrome Extension

    Two hours of free runtime, no credit card required