scrapeprices Who changed their prices.
Pricing pages get edited quietly. Every re-read of a provider's own page is diffed against the last one, and anything that moved lands here with the date and the source.
Subscribe by RSS- method
Sellers who price by target site are now ranked on a published basket of ten real sites
Four sellers here do not have one price per page: Zyte, ScraperAPI, Scrape.do and Crawlbase assign the price from the site you point it at, and Oxylabs charges Amazon and Google their own rates. Ranking them on the floor of their own range put a best case at the top of the table beside flat rates that apply to any URL, and ranking them on the ceiling would have punished them for publishing a range at all. From today the table looks up each seller's own published price for each of ten fixed target sites - Amazon, eBay, Walmart, Google, LinkedIn, Indeed, Zillow, Yelp, Instagram and TikTok - and ranks on the median. A flat rate prices all ten, so its median is its rate card and its position does not move. A target a seller refuses, or prices only behind a sign-in, is left out of that seller's median and named on its page. The basket is listed in full on the methodology page and in prices.json.
Source - price up Z
Zyte API Zyte's median real page is $0.57 per 1,000, not the $0.13 headline
Zyte's own website-pricing estimator answers without an account, so its tier for a named site is public. Across the ten basket sites it puts Amazon and eBay in tier 2 Easy, Zillow in tier 3 Moderate, and Walmart, Indeed and Yelp in tier 4 Complex. Not one came back tier 1, which is the $0.13 the pricing page leads with. The median plain page moves from $0.13 to $0.57 per 1,000 and the median rendered page from $1.01 to $2.01. Its anti-bot figure moves the other way, from $1.27 to $0.57, because this site had been modelling anti-bot as Zyte's dearest plain tier; Zyte sells no separate stealth tier at all, so the tier the site lands in is the anti-bot price.
Source - gone Z Zyte API
Zyte will not scrape LinkedIn at any price
Typed into Zyte's own pricing estimator, linkedin.com comes back with no tier and no price where every priced site shows one, and the estimator's data marks it unavailable. Google, Instagram and TikTok come back marked behind a login, also with no price. Zyte is not dearer on those four; it does not sell them.
Source - price up S
ScraperAPI ScraperAPI names the hosts, and three of them are the commerce sites buyers came for
Its credits table prints the domain class with the host names inside it: Amazon, Walmart and eBay cost 5 credits before any parameter, Google and Bing 25, LinkedIn 30, against a 1-credit flat rate. Its own per-class grid supplies the rendered and residential figures, published under the line that DataPipeline credit costs now match its standard scraping rates. The median plain page moves from $0.49 to $1.49 per 1,000 at 100,000 a month. Five of the ten basket sites it declines to price at all: an unnamed host is 1 credit unless it sits behind Cloudflare, Turnstile, DataDome or PerimeterX, which is 10, and ScraperAPI publishes no list of which sites those are.
Source - price up C
Crawlbase Crawlbase does publish the difficulty tier, in a third place nobody looks
This site had recorded Crawlbase's per-domain tier as unknowable. That was wrong: its Scraping Cookbook carries a page for each of 383 hosts naming the tier and the credit cost, and six of the ten basket sites have one. LinkedIn is Extreme II at 20 credits, Walmart Complex II at 4.5, Google Complex at 2, and Amazon, Yelp and TikTok Moderate at 1.5. The median plain page moves from $0.60 to $0.88 per 1,000 at 100,000 a month, and Crawlbase now carries an anti-bot price where it had a blank, because residential exit nodes and Cloudflare handling carry no surcharge and the tier is the whole bill. Rendering is the tier doubled, which Crawlbase states rather than leaves to arithmetic: "Rendering in a real browser doubles the credits of any request, whatever the site tier." Two of its own surfaces still fall short of the Cookbook: the API docs say the complexity is "one of standard, moderate, or complex" and send you to sales for the price, and the cost estimator at the top of the pricing page offers only x1, x1.5 and x2, though the full ladder is printed further down the same page, so sizing a LinkedIn job on that widget under-budgets it tenfold.
Source -
Oxylabs rejects LinkedIn with an HTTP 400 before it scrapes anything
Its restricted-target list names LinkedIn under professional and social networks: "A request to a restricted target is rejected with HTTP 400 before any scraping happens." Access is possible only after a KYC conversation with support, which is not a published price, so LinkedIn is left out of its basket median. Nothing named linkedin exists anywhere else on oxylabs.io. Amazon and Google keep their own rows on every plan card, but both discounts vanish the moment you ask for JavaScript, because the rendered rate is a single figure per plan.
Source - added
First capture: 26 web scraping APIs
Every provider's published plan table read off its own pricing page and normalized to cost per 1,000 pages at four monthly volumes, for each of the five ways a page gets fetched: plain, rendered in a browser, through a residential address, through an anti-bot wall, and returned as Markdown. From here, each pricing page is re-read and any change is logged on this page with the before and after.
-
NetNut's domain has been seized by the FBI
Every netnut.io URL, including the home page, the pricing page and the Website Unblocker page, now returns a single page titled "Seized by the Federal Bureau of Investigation". There is no published price to read and no product to buy. The entry stays listed with nothing in it, so that anyone searching for NetNut learns that rather than nothing.
Source - gone S
ScraperBox ScraperBox is gone, and its domain now serves a gambling site
scraperbox.com answers, but what it serves is an Indonesian gambling site rather than a scraping API. The pricing page and the docs both 404, and api.scraperbox.com no longer resolves in DNS. It stays listed with no price for the same reason as any other dead entry.
Source
