Extract clean content automatically with Zyte API’s new pageContent data type

Today, we’re opening access to pageContent, a brand-new Zyte API automatic extraction data type that delivers the content of any web page as plain text minus the noise.

Everyone who carries out web data extraction is ultimately focused on the same end goal – obtaining the main content of a destination page. But the need to target underlying CSS or XPath selectors for that content complicates the life of every data engineer. 

With pageContent, instead of targeting fragile selectors that are prone to changing, you just send a URL and get back the content from the main body of the page. It’s the fastest route to extract clean LLM friendly content from web pages as plain text.

But pageContent doesn’t only return a page’s main content. It also automatically returns smart navigation links and pagination links, without you needing to script any parsing logic.

Think of pageContent as a one-call content fetcher: URLs go in; clean, structured data comes out. 

While a regular Zyte API request returns a target page’s HTML, passing pageContent: true will strip out all the noise and return just the text you want wrapped in JSON with some useful extras like the Xpath for the main content, links and some meta.

itemMain can return the text of a news article, the description of a product listing, the content of a forum post – pretty much, whatever is the main text material of a page.

pageContent is clever and helpful enough to identify and return all these all these useful elements from a target page, to better support your scraping workflows:

All of this arrives as clean JSON, ready to use in your applications.

pageContent is a close cousin of the other data types that Zyte API can already return automatically through content-specific schemas in structured JSON, thanks to the supervised machine learning models underpinning our automatic extraction:

• Product

  • product

  • productList

  • productNavigation

• Articles 

  • article

  • articleList

  • articleNavigation

• Forum

  • forumThread

• Job Listing

  • jobPosting

  • jobPostingNavigation

• Search Engine Results

  • SERP Organic listings (free to use at time of writing)

If you’re tired of writing code, and sick of selectors breaking, automatic extraction is the quickest and easiest way to cut to the chase – AI handles the job of turning HTML into structured data.

But, while using existing automatic extraction data types depends on users already knowing the content type of the end content they want to extract, pageContent does not need this input; it will happily return the page’s main content and other standard elements, whatever the type.

Because pageContent runs through Zyte API, you also get Zyte’s advanced unblocking, proxy rotation and headless browser tech included.

Manual scraping can be painful. The average web page carries a lot of noise for every signal Developers spend hours inspecting ever‑shifting DOMs and updating spiders when a div class changes.

• Just get results: pageContent lets you skip the pain of parsing. At the end of the day, scraping exists to deliver results, and pageContent makes that quicker, easier and more easily scalable.

• Light on schema, infinite content: Existing automatic parsing tools on the market return a page’s intrinsic content in a content-specific schema, like article or product description. pageContent returns content in the same schema every time, so your extraction workflow is highly flexible.

• Cut LLM token costs. Feeding an entire HTML document to a large language model (LLM) is often expensive. itemMain removes boilerplate, nav menus and ads from scraped content, so you can save token and processing costs by supplying only meaningful inputs to your AI.

We think pageContent is going to be pretty popular. For instance:

1. LLM ingestion – Summarise, classify or translate fresh web pages without paying for useless markup.

2. Content monitoring and archiving – Store the main text and ignore the rest.

3. SEO / accessibility audits – Validate presence of headings, breadcrumb trails and pagination in one response.

4. Crawl orchestration – Use the nextPage URL to drive spiders, zero regular expressions required.

Any time you need pre-parsed, clean text extracted from web pages at any scale, pageContent is going to be a massive time saver, both upfront and ongoing.

Using pageContent is as simple as enabling pageContent in a JSON request body sent to Zyte API’s /extract endpoint.

You can read the docs and try the same call inside the Zyte API Playground.

pageContent unlocks new extraction efficiencies for all Zyte’s customers:

• Developers and data engineers — no more selector maintenance; lower compute bills.

• Technology leaders — scalable, compliant data pipeline without new infrastructure or headcount.

Product / business intelligence leads — faster time‑to‑data; predictable costs; API drops data straight into pipelines.

pageContent is live in Zyte API now, and the docs are available to read Try it out in the Zyte API Playground and see how clean content looks when the boilerplate disappears.

👉 Start your first pageContent request now

Questions? Jump into our Discord community — we love hearing what you build when the page just melts away.