What is the cheapest way to scrape a million pages a month?
A million pages a month costs from $90.00 as plain HTML on ScrapingDog ($0.09 per 1,000), from $350.00 when they need JavaScript on ScrapingDog, and from $699.00 on sites that block bots, on ScrapeBadger, at a 5% failure rate.
Worked out from each vendor's own pricing page; each vendor page on this site shows the day its price was read. 5% of requests are assumed to fail; bills that rest on an assumption are left out.| Page type | API | Monthly bill | Per 1,000 pages | Plan |
|---|---|---|---|---|
| Plain HTML | ScrapingDog | $90.00 | $0.09 | Standard |
| Scrape.do | $99.00 | $0.10 | Pro | |
| ScrapeOps | $99.00 | $0.10 | 1,250,000 credits | |
| ScrapeOwl | $99.00 | $0.10 | Startup | |
| ScrapingBee | $99.00 | $0.10 | Startup | |
| With JavaScript | ScrapingDog | $350.00 | $0.35 | Premium |
| Olostep | $399.00 | $0.40 | Scale | |
| ZenRows | $549.00 | $0.55 | Scale S1 | |
| ScrapingBee | $599.00 | $0.60 | Business + | |
| ScrapingBot | $599.99 | $0.60 | Business | |
| Sites that block bots | ScrapeBadger | $699.00 | $0.70 | Scale |
| Scrape.do | $699.00 | $0.70 | Advanced | |
| HasData | $830.00 | $0.83 | Growth 10M | |
| WebScraping.AI | $1,999.60 | $2.00 | Pay as you go |
12 of 40 web scraping APIs publish a complete bill for a million JavaScript pages; the rest sell that volume by quote or leave part of the price unpublished. At this volume an enterprise quote can beat the list price, so the figures above are the price to beat.
Why the page type matters more than the vendor
At a million pages the gap between page types is larger than the gap between vendors. A plain HTML page is the vendor’s base request; a page that needs a browser to render costs several times that on most APIs, and a site that blocks bots needs premium or residential proxies or a stealth mode, priced higher again. Before comparing vendors, sample a hundred of your target pages and count how many truly need JavaScript and how many get blocked: a job that is 80% plain HTML costs a fraction of one priced as all rendered pages.
What the figures include
Each bill is the vendor’s own published plan or rate for 1,000,000 pages a month, with 5% of requests assumed to fail and billed or not as each vendor says. Bills that rest on an assumption about an unpublished rate, one-off prepaid packs and older readings are left out of the ranking. Vendors that sell this volume only by quote are not in the table; at this size their quote is worth asking for, and the figures here are the price it has to beat.
To price your own mix, enter 1,000,000 on the calculator and switch JavaScript and protected sites on and off; the cheapest scraping API page explains how a “cheapest” claim is made.