Is web scraping free?
The code is free; pages at any real volume usually are not. Scrapy and Beautiful Soup are free, open-source libraries, and a scraper run from your own connection costs only the machine it runs on, until sites start to block it or need a browser to render. A scraping API's free tier covers the first pages: the largest that renew give 10,000 plain pages a month (Decodo Web Scraping API and ScrapingAnt) and 6,000 with JavaScript (Decodo Web Scraping API). Past the free tiers, 100,000 JavaScript pages a month cost from $49.00 on ScrapeBadger.
Worked out from each vendor's own pricing page; each vendor page on this site shows the day its price was read. Free pages are each allowance divided by the vendor's own rate for a plain page and for a JavaScript page.| API | Free plain pages a month | With JavaScript | Card | After the free tier |
|---|---|---|---|---|
| Decodo Web Scraping API These counts use the standard proxy pool ("proxy_pool": "standard"). Left at Decodo's default premium pool, each page bills double, so the free plan covers 5,000 pages. | 10,000 | 6,000 | not needed | $19 a month |
| ScrapingAnt | 10,000 | 1,000 | not needed | $19 a month for 100,000 credits |
| ZenRows | 5,000 | 1,000 | not published | $19 a month for 45,000 credits |
| Bright Data Web Unlocker and Web Scraper API | 5,000 | no rendering rate published | not needed | pay as you go at $1.50 per 1,000 requests; plans from $499 a month |
| WebScraping.AI | 2,000 | 400 | not needed | pay as you go at $0.20 per 1,000 credits; plans from $29 a month |
| Context.dev | 1,000 | 1,000 | not needed | $19 a month for 7,500 credits |
| Firecrawl | 1,000 | 1,000 | not needed | $19 a month for 5,000 credits |
| Scrape.do | 1,000 | 200 | not needed | $29 a month for 250,000 credits |
| HasData | 1,000 | 100 | not needed | $59 a month for 200,000 credits |
| BrowserCloud | 1,000 | no rendering rate published | not published | $19 a month for 120,000 credits |
A site that blocks bots costs several times a plain page through an API's premium proxies, so a free allowance buys far fewer of those pages. Running your own scraper at volume moves the cost to proxies and servers instead; the proxies tool compares that route with an API for your volume.
What costs nothing
The software. Scrapy calls itself “free, open source” (Scrapy, checked Oct 12, 2026), and Beautiful Soup is released under the MIT license (Beautiful Soup, checked Oct 12, 2026). A script that reads a few hundred pages a day from sites that do not mind, over your own connection, costs nothing beyond the computer it runs on.
What starts to cost money
Blocking. A site that limits bots answers a busy address with a CAPTCHA, a 403 or a 429. Getting past it takes rotating proxies, paid by the gigabyte or by the IP, or an API that rotates them for you. On a credit-based API a protected page costs about ten times a plain one: Scrape.do and ScrapingDog both charge 1 credit for a plain request and 10 with residential or premium proxies (Scrape.do request costs, ScrapingDog pricing, checked Oct 12, 2026).
JavaScript. A page that builds itself in the browser needs a headless browser to render it. Running one yourself takes far more memory and time per page than a plain request; on the same two APIs a rendered page costs 5 credits, and 25 with premium proxies as well.
Volume and upkeep. At tens of thousands of pages a month the machines, the proxies and the time spent fixing a scraper after a site changes are the real bill. An API folds the first two into one price per page, which is why the table counts free tiers in pages rather than credits.
When a free tier is enough
A free tier that renews every month can carry a small recurring job, such as a price check on a few hundred products a day. The table lists the largest; one-off sign-up credits and timed trials are not in it, because they run out once. Every free tier, renewing or not, is on the free tiers page, and the proxies tool compares running your own scraper through proxies with paying an API at your volume.