We use affiliate links. They let us sustain ourselves at no cost to you.

Zyte Adds AI Scraping Functionality to Its API

The tool can now crawl, unblock, and parse websites using AI and an optional no-code interface.

Adam Dubois
zyte ai scraping promo
Image source: Zyte
Zyte, the Ireland-based web data extraction platform, has announced the addition of new AI-based functionality to Zyte API. With it, the tool has become a “complete solution for web scraping”, allowing developers to “build new spiders and add data sources in minutes”.  The solution combines Zyte API and pre-made spider templates accessible through Scrapy Cloud on Zyte’s dashboard. It allows extracting data from websites without creating your own crawling logic, dealing with anti-bot systems, or specifying selectors for data extraction.   The data extraction step relies on Zyte’s proprietary machine learning model that the company claims to be up to “50x cheaper and 56% more accurate” than large language models like ChatGPT. For now, the only available template covers e-commerce product pages. Using Scrapy Cloud’s no-code interface, you can quickly create a web scraper by specifying a few parameters: starting URL, geolocation, request count, crawling strategy, and browser rendering. The scraper then automatically crawls the website and returns structured data. Developers who need more functionality can make quick customizations on the dashboard or use the provided Python library. This way, it’s also possible to build completely new scrapers that make use of Zyte’s AI tech.

The company first introduced its e-commerce template and no-code interface during Extract Summit, Zyte’s annual web scraping conference (you can find our recap here). We’ve tried it several times for small tasks like crawling e-commerce pages and received satisfactory results.

All in all, Zyte’s AI Scraping is a fascinating project that tries to combine the interests of both business folks and seasoned Scrapy developers. For now, the experience is still clunky: it requires two subscriptions (to Zyte API and Scrapy Cloud), parses limited data types, and the user interface can be intimidating. In addition, the API’s pricing isn’t always easy to estimate, as it changes dynamically based on the target. 

Having said that, the underlying tech is, without a doubt, solid, and we’re eager to see how the provider will streamline its product in the future. 

You can try the AI e-commerce template for free by claiming Zyte API’s $5 credit and using Scrapy Cloud’s free tier.

Get proxy news and updates directly to your inbox.



Picture of Adam Dubois
Adam Dubois
Proxy geek and developer.
More by this author