scrapeprices JSON

How a price on this site is calculated.

Scraping pricing pages are not comparable to each other. They quote credits, requests, successful responses and gigabytes; they hide a minimum in a footnote; they price the same page at one credit or seventy-five depending on how hard it was to get, and on several price lists the same page has no fixed price at all. Six steps turn all of that into one number.

  1. 1Read the price off the provider’s own page

    Every plan tier, credit rate and minimum comes from a page the provider publishes. Where a page is JavaScript-only and unreadable, the source is recorded as whatever we did read, on the provider record. A number nobody publishes is left blank rather than estimated.

  2. 2Convert credits into pages

    A credit is not a page. Nearly every seller here bills one URL as some number of credits, and the number moves with how the page was fetched: plain, rendered in a browser, through a residential address, or through a stealth pool built for anti-bot walls. Each record carries the multiplier the provider prints for each of those, and the table applies it before comparing anything. Where a provider sells a tier but never prints its multiplier, the cell says so rather than assuming one.

  3. 3Price the sellers whose price depends on the target

    Some sellers here do not have one price per page. The same fetch costs one figure on an easy site and many times that on a hard one, and the seller decides which is which. Ranking those on the floor of their own range would put a best case at the top of a table beside rates that apply to any URL, and ranking them on the ceiling would punish them for publishing a range at all. So this site names 10 real targets, looks up each seller's own published price for each of them, and ranks on the median. A seller with one flat rate charges it for all 10, 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 said out loud on its page.

  4. 4Buy the volume the cheapest published way

    For each of the four monthly volumes we take the cheapest route the provider actually sells: the smallest plan that covers it, or a larger plan plus published overage, or the pay-as-you-go rate. Routes are chosen on what a year costs, and a deposit or pack a buyer would not use up in that year counts as spent, so a $500 top-up never wins a column for a $6-a-month workload. The monthly figure is then one month of usage at the chosen rate. A plan with a hard cap and no overage rate cannot serve a volume above its allowance, so it is not considered for that column. Queued delivery, where results arrive minutes later, is priced separately and shown beside the live price, never in place of it.

  5. 5Charge the plan floor, not the sticker

    If the cheapest way to buy 1,000 pages a month is a $50 monthly plan, the price is $50 per 1,000, not the per-page rate printed next to it. Any monthly minimum is applied on top. A prepaid pack is different: its credits carry across months, so a month costs the pages used at the pack rate and the pack price is shown as the up-front gate. When a pack would die with credits unused, the month is charged the pack spread over its validity, because that money is spent either way.

  6. 6Rank by price, with no exceptions

    Rows sort on the computed per-1,000 figure at the selected volume. Nothing is boosted, pinned or paid for. Providers we could not price at all are listed under the table so their absence is visible rather than silent.

The four volumes, and why

One price per provider is meaningless, because the ranking inverts as volume rises. A pay-as-you-go provider with no floor wins at 1k a month and loses at 1M; a provider with a $500 commitment is the reverse. So every table here carries all four, and you pick the one that matches your workload.

1k
pages per month

A side project, a research script, one client’s rank check.

10k
pages per month

A working product with a handful of customers.

100k
pages per month

A rank tracker or lead tool at real scale.

1M
pages per month

A data business. Here the per-call rate is all that matters.

The 10 sites every seller is priced against

Ten sites this market is actually bought for: three of the big commerce catalogues, search, jobs, real estate, local business data, two social networks, and the one target that half the sellers refuse. The list is fixed and published so any reader can repeat the lookup, and it is the same ten for every seller.

Target Why it is in the basket
amazon.com the most-scraped commerce catalogue in the market
ebay.com the second commerce catalogue, and historically an easier one
walmart.com the third, and historically a harder one
google.com search results, the single largest scraping workload there is
linkedin.com the hardest mainstream target, and the one sellers refuse outright
indeed.com jobs, a large vertical every seller in this table addresses
zillow.com real estate, a long-standing anti-bot target
yelp.com local business data
instagram.com social, behind a login wall
tiktok.com social, with the most aggressive bot defences of the ten

What this table cannot tell you

Corrections

26 providers are tracked and prices were last checked on 2026-09-15. If a number is wrong, send the provider page that shows the right one to hello@scrapeprices.com. Corrections are made against the source, not against an argument, and the change shows up in the changelog. Providers are welcome to ask for a re-read; nobody can buy a position.