Zyte Pricing and Real Cost per 1,000 Pages Across Zyte API Tiers

Zyte is the company behind Scrapy, and its pricing reflects an engineering culture: honest, granular and published in full. It is also the hardest pricing in this market to turn into a monthly number before you have run anything, because the rate you pay is set per website by Zyte's own classifier rather than chosen by you.

This piece does the conversion. Every rate below was read from zyte.com/pricing and the Zyte API pricing documentation on 26 September 2026. Zyte reviews its tier assignments on a schedule, so check the live pages before committing a budget.

The Zyte API rate card, per 1,000 successful requests

There are two request types and five tiers of each. HTTP requests return the raw response body. Browser requests render the page in a headless browser first, which is what any JavaScript heavy site needs. These are the pay as you go rates.

Two things stand out. The HTTP column is genuinely cheap: tier 1 at 0.13 USD per 1,000 is among the lowest published rates anywhere for a managed fetch. And the browser column doubles at every step, so a site that moves from tier 3 to tier 4 doubles your rendering bill overnight without a line of your code changing.

Extras are billed on top: screenshots at 0.002 USD each, which is 2 USD per 1,000, and automatic extraction at 0.0004 to 0.0016 USD per data type per request. If you want parsed product fields rather than HTML, add 0.40 to 1.60 USD per 1,000 pages to every row above.

Who picks your tier is the detail third party reviews skip

Every review of Zyte we read restates the range. None of them mentions who decides where in the range you land. Zyte's documentation is explicit: tier assignment is automatic, new site and request type combinations start on a temporary tier until Zyte has gathered enough data, and assignments are reviewed quarterly with two weeks' notice of changes.

In practice that means three things for a buyer. You cannot price a new target from the rate card, only from a test run. The rate for a target you already scrape can move each quarter, in either direction, as the site's defenses change. And a crawl across many domains lands on a blend of up to ten rates that you only see after the fact. Zyte's own dashboard shows the tier per domain, and it is worth reading before you sign any commitment.

That is not a criticism of the engineering. Charging more for harder sites is fair. It is a forecasting problem, and forecasting is what the person approving the budget cares about.

The 100 USD ceiling on pay as you go

Pay as you go carries a 100 USD monthly spending limit. That sounds generous until you divide it by the browser column.

So pay as you go is really a trial for anyone rendering protected sites. Around 25,000 moderate rendered pages a month and you are on a commitment plan whether you planned to be or not. New accounts also get 5 USD of credit valid for 30 days, which is enough for several thousand HTTP requests or about 300 of the hardest rendered ones, and that is the right way to find out which tiers your targets fall into.

Commitment plans and what they really cost

Committing to a monthly minimum buys a discount across every tier and raises the spending limit. The published ranges work out to these discounts off pay as you go.

Zyte's documentation also lists a 350 USD commitment between the last two, at about 48 percent off with a 700 USD spending limit. The commitment is paid at the start of each billing month. Use more and the overage lands on the next bill. Use less and, in Zyte's words, no refund is provided. The top standard plan allows up to 2,500 USD of spend a month, and above that you are talking to sales about Enterprise terms.

The practical trap is the step between plans. If your usage sits at 180 USD of pay as you go value, the 200 USD plan's 40 percent discount brings it to about 108 USD, but you still pay the 200 USD minimum. Commit only to the level your lowest month will clear.

Real budgets, Zyte against flat per request pricing

The honest comparison depends almost entirely on whether your pages need a browser. Here are five realistic monthly workloads, priced at the cheapest Zyte plan that fits inside its spending limit, against our flat rate of 0.98 USD per 1,000 on Starter, 0.60 on Growth and 0.33 on Scale, which does not change with the target.

The pattern is clean. On plain HTTP pages Zyte is cheaper than we are, often by a lot, and we would not argue otherwise. Once a browser is involved the picture flips, and on hard targets it flips by a factor of ten. The breakeven sits around tier 3: even at the deepest published discount, a rendered page on Zyte costs more than our Starter rate from tier 3 upward and more than our Scale rate at every browser tier.

Where Zyte is the right buy

High volume, unrendered crawling of ordinary sites. If your targets serve their content in the HTML, as a surprising number of catalog and listing sites still do, Zyte's tier 1 and 2 HTTP rates are hard to beat, and success based billing means rate limited and failed responses cost nothing.

Teams already on Scrapy. The Scrapy integration is first party, the maintainers work there, and Scrapy Cloud gives you hosted spiders and scheduling in the same account. If your crawlers are Scrapy projects, that fit is worth real money in engineering time.

And the rate card is published per tier rather than hidden behind a credit multiplier table, which puts Zyte ahead of several vendors we have priced in this series. The 3,000 requests per minute limit on standard plans is also generous for the money.

Where flat per request pricing wins

Rendered pages and protected targets, which is where most of the money in scraping is spent. A retail or travel site behind Cloudflare, Akamai or Imperva and Kasada is almost by definition a tier 4 or 5 browser target on a per tier rate card. On a flat rate it costs the same as a static blog, and refused attempts are not billed.

It also wins when somebody has to approve the budget. A flat rate turns next quarter's scraping cost into pages times a number, with no classifier review that can move it, no commitment minimum to clear and no ceiling to trip halfway through the month. If you run crawls on a schedule, put an uptime check on the webhook endpoint that receives the results, too. The usual cause of a missing day of data is the receiving end going down, not the fetch.

If your workload is a mix, which most are, run both numbers on your real target list. Zyte's 5 USD credit and our plans make that a one day exercise.

Questions buyers ask

How much does Zyte cost? Zyte API is billed per 1,000 successful requests, from 0.13 USD for simple HTTP pages to 16.08 USD for the hardest browser rendered pages on pay as you go. Monthly commitments of 100, 200 or 500 USD cut those rates by about 25, 40 and 52 percent. Screenshots and automatic extraction are billed on top.

Does Zyte have a free trial? New accounts get 5 USD of credit for 30 days with no commitment. That is enough to learn which tiers your target sites fall into, which is the most useful thing to do with it, because you cannot know your real rate from the pricing page alone.

Does Zyte charge for failed requests? No. Zyte bills only successful responses, and rate limited and unsuccessful responses are free. The same is true of ScraperAPI, Oxylabs and us, so it is table stakes in this market rather than a reason to choose any one vendor.

How are Zyte API tiers assigned? Automatically, per website and request type. New combinations start on a temporary tier until Zyte has enough data, and assignments are reviewed quarterly with two weeks' notice of changes. Your cost for a given site can therefore move between quarters without any change on your side.

Is Zyte cheaper than ScraperAPI? For plain HTTP pages, usually yes. ScraperAPI's entry plan works out to 0.49 USD per 1,000 normal pages and several times that once JavaScript rendering and protected target credits stack up. We converted both to the same unit, so compare this page with our ScraperAPI pricing breakdown, and the Oxylabs, Bright Data, Apify and ScrapingBee conversions alongside it.

Is Zyte good for scraping Google results? Its rates are low for plain HTTP fetches, but a parsed result page with positions and features is a different product. If rankings are the job, a dedicated rank tracking API that bills one request per results page is the simpler comparison.