What a scraping API credit buys: the rendering and proxy multipliers, vendor by vendor
A plan of 100,000 credits is 100,000 plain pages at most vendors, 20,000 rendered ones, and far fewer on protected sites. The multipliers, from the vendors' own tables.
Published 2026-09-18 by Phillip Paul Inker. Sources re-read 2026-09-18. 8 sources.
The plan counts credits, your targets decide what they buy
Every credit-based vendor here sells the same headline: so many credits a month for so many dollars. The number that matters is the second one on the documentation page, the cost of a request once it needs a browser, a better proxy, or a named hard domain. ScrapingBee states it plainly: it costs 1 to 75 credits to make a request, depending on the parameters. A plain fetch is one credit, the default headless-browser fetch five, premium proxies ten, both together twenty-five, stealth proxies seventy-five.
The same shape repeats with different numbers. ZenRows runs a 250,000-credit balance as about 250,000 standard pages, 50,000 with JavaScript or 10,000 hard-protected, so 1, 5 and 25. Scrapfly is 1 for an HTTP scrape on a datacenter IP, 5 with JavaScript rendering or its unblocker, 25 with residential proxies, 60 for a full-page screenshot. Scrapingdog is 5 for rendering, 10 for a premium proxy, 25 for both. Scrape.do is 5 for a datacenter request with the headless browser, 10 for a residential or mobile proxy, 25 for both. ScraperAPI is the outlier at 10 for rendering, 25 with premium proxies, and 75 for ultra premium with rendering. Crawlbase is the simplest: a real browser render doubles the request, and that is the whole story.
What that does to the price per 1,000
Take the entry plan of each vendor and divide: ScrapingBee's $19 for 75,000 credits is $0.25 per 1,000 plain requests and $1.27 per 1,000 rendered. ScraperAPI's $49 for 100,000 is $0.49 plain and $4.90 rendered. ZenRows' $16 for 45,000 is $0.36 plain and $1.78 rendered. Scrapfly's $30 for 200,000 is $0.15 plain and $0.75 rendered. Scrapingdog's $40 for 200,000 is $0.20 plain and $1.00 rendered. Scrape.do prints $0.11 per 1,000 on its $29 card, so $0.55 rendered. Crawlbase's $99 block of 200,000 is about $0.50 plain and $1.00 rendered.
Two vendors skip the arithmetic. Decodo prints the per-1,000 price for each mode on each plan, $0.50 standard and $0.75 with JavaScript on its $19 plan. Firecrawl charges one credit per page because the rendered, cleaned page is its unit; its multiplier is the plan size, 5,000 pages for $16 at annual prices.
Failed requests
Every vendor in this database that sells credits says a failed request costs nothing, in its own words: ScrapingBee pays only for successful requests, Scrapfly never charges on error, Scrapingdog never charges failed requests, Scrape.do charges only on status 2xx, 400, 404 and 410, Crawlbase never bills a request that fails or is blocked. Read the definition of failure, though: a 200 with a CAPTCHA page in the body is a success to a billing system.
- Find the credit table in the documentation before the plan page
- Divide the plan price by credits, then by the multiplier your targets need
- Send one request per target and read the cost header
- Check whether the free tier is monthly (ZenRows, ScrapingAnt, Firecrawl, Scrape.do) or once (ScrapingBee, Scrapfly, Scrapingdog)
- [1]ScrapingBee documentation, credit costs ScrapingBee, read 2026-09-18.
“It costs 1 to 75 credits to make requests to the ScrapingBee API. By default, ScrapingBee fetches the URL to scrape via a headless browser ... costs 5 credits per request.”
- [2]ZenRows pricing ZenRows, read 2026-09-18.
“Runs about 250K standard pages, 50K with JavaScript, or 10K hard-protected, all from one shared balance.”
- [3]Scrapfly pricing, what a credit buys Scrapfly, read 2026-09-18.
“1 credit | HTTP scrape on datacenter IP. 5 credits | + JS rendering or Unblocker. 25 credits | + residential proxies. 60 credits | screenshot capture (full-page). Failed request | $0, never charged on error.”
- [4]Scrapingdog documentation, credit cost of APIs Scrapingdog, read 2026-09-18.
“Web Scraping — JS rendering | 5. Web Scraping — JS rendering + premium proxy | 25. Web Scraping — premium proxy | 10.”
- [5]Scrape.do documentation, request costs Scrape.do, read 2026-09-18.
“Normal request (datacenter proxy) 1. Datacenter + headless browser (render=true) 5. Residential / mobile + headless browser (super + render) 25. google.* 10. aircanada.com 75. jd.com 75.”
- [6]ScraperAPI documentation, credits and requests ScraperAPI, read 2026-09-18.
“render=true - 10 API Credits/req. premium=true + render=true - 25 API Credits/req. ultra_premium=true + render=true - 75 API Credits/req.”
- [7]Crawlbase pricing Crawlbase, read 2026-09-18.
“A real browser render doubles the request, and that is the whole story. There is no browser tier. A request that fails or is blocked is never billed.”
- [8]Decodo Web Scraping API price table Decodo, read 2026-09-18.
“All prices shown are per 1K req. $19: $0.50 $0.75 $1.00 $1.50 10 req/s.”
Reviews this guide draws on
ScrapingBeesubscriptionThe credit table is the review: a plain request is one credit, the default headless-browser request five, premium proxies ten, both together twenty-five, stealth proxies seventy-five. Cheap for plain fetches, ordinary once JavaScript and premium proxies are on.
ScraperAPIsubscriptionSeven plans with threads and geotargeting scope printed on each, a credit-cost page that names the parameters and the domains that cost more, and a written 7-day no-questions refund. Rendering at ten credits makes it dearer than most for JavaScript pages.
ZenRowssubscriptionA clear ladder: five paid tiers from $16 to $456, concurrency printed on each, rollover on paid plans, and the credit rule stated in one sentence. What is missing is any success or latency figure; the word on the page is reliable, not a number.
ScrapflysubscriptionFeature access never gates, the page says, and the plans bear it out: every tier has all five APIs, rendering, the unblocker, residential proxies in 190+ countries and the SDKs; only volume, concurrency, log retention and support change. The credit table is four lines and failures are free.
Scrape.dosubscriptionThe most legible pricing in this database: the plan card prints the price per 1,000 credits, the documentation prints what each request type and each named domain costs, and only successful requests count. A 99.98% success figure on the homepage is a claim, not a method.
ScrapingdogsubscriptionA long, plain ladder: credits and concurrency printed on every tier, JavaScript rendering at five credits and premium proxies at ten in the documentation, and every dedicated API included on every plan. The free tier is a test drive of 200 credits, not a working allowance.
Crawlbaseper requestPricing written like a tax table: pay as you go on a ladder of brackets, or a monthly block at a lower effective rate with the ladder continuing past it. A failed or blocked request is never billed and a rendered one costs two credits, which is the simplest multiplier in this database.
Run a scraping API?Get it on the record.
Send us the pages that state your prices, credit costs and limits. We read them, score them on the public rubric, and publish the review with the date.