Scraping API PricesHow much does it cost to scrape 100,000 pages?
On ScrapeOps, 100,000 successful normal-site pages cost $19 per month without JavaScript, $54 with its limited-provider render_js_cheap mode, or $99 with standard JavaScript rendering. As of July 26, 2026, source: ScrapeOps plans and request-cost schedule, with meter details from ScrapeOps documentation, retrieved July 26, 2026.
The exact arithmetic
A normal successful response costs 1 credit, so 100,000 pages consume 100,000 credits and fit the $19 monthly plan. The lower-cost JavaScript mode consumes 5 credits per response, so the same page count consumes 500,000 credits and fits the $54 plan. Standard JavaScript rendering consumes 10 credits per response, so the workload consumes 1,000,000 credits and fits the $99 plan. ScrapeOps plans and request costs, retrieved July 26, 2026.
Those are three different workloads, not three prices for equivalent service. ScrapeOps labels render_js_cheap as a lower-cost mode with a limited provider set, while render_js=true is the standard rendering mode. ScrapeOps request costs, retrieved July 26, 2026.
The calculation also uses the vendor’s billing definition of a successful page. ScrapeOps consumes credits for HTTP 200 and 404 responses, and unused credits do not transfer to the next subscription period. A returned 404 can therefore be billable even when it is unusable for the buyer’s project. ScrapeOps request costs, retrieved July 26, 2026.
Named targets can replace the generic rate. ScrapeOps publishes 10 credits for Google without rendering, 25 for Google with rendering, 35 for G2 and Leboncoin, and 70 for LinkedIn; its anti-bot modes range from 10 to 80 credits per request. A 100,000-URL file containing those targets is not the normal-site workload priced above. ScrapeOps domain and bypass costs, retrieved July 26, 2026.
Residential access and extraction also change the total. Residential plus JavaScript costs 25 credits per successful request, while LLM extraction adds 25 credits to the selected base mode. ScrapeOps request costs, retrieved July 26, 2026.
The page count should include pagination and every client retry sent as a new request. It should not include an invented vendor success rate, because this site has not benchmarked providers; use the vendor’s observed cost header and your own accepted-output count after a sample run. Scrape.do and ScrapingAnt both expose a response header containing the actual credit cost, providing a first-party pattern for validating request classification before scale. Scrape.do request costs and ScrapingAnt credit documentation, retrieved July 26, 2026.