Skip to content
proxycosts
Prices checked up to Nov 17, 2026

How much bandwidth does web scraping use?

Priced, checked up to Nov 17, 2026:

Every billing type competes; free plans are left out. Change the amount on each comparison.

Pages × size per page × retries. Fetching only the HTML, a typical page is about 33 KB, so 200,000 pages with 20% retries come to about 7.9 GB. Loading whole pages in a headless browser, with images and scripts, a typical page is about 2.6 MB, so 10,000 pages with the same retries come to about 32 GB. Measure a sample of your own targets to replace these medians; the box above prices 10, 50 and 100 GB of residential traffic today.

Answer · The comparison uses prices checked up to Nov 17, 2026; every figure in the text keeps the date it was checked.

On this page

Proxies bill for every byte that passes through them, so the traffic a scrape uses is set by what your scraper downloads, not by the number of pages alone. The same 10,000 pages can cost a few hundred megabytes or tens of gigabytes depending on how you fetch them.

The arithmetic

monthly traffic = pages a month × size per page × (1 + share of requests retried)

Job Pages a month Size per page Retries Traffic a month
Product pages, HTML only 200,000 33 KB 20% 7.92 GB
Search result pages, HTML only 1,000,000 33 KB 30% 42.9 GB
Pages in a headless browser 10,000 2,652 KB 20% 31.8 GB

The sizes are medians from the HTTP Archive’s Web Almanac 2024 (the HTML of a page) and its page weight chapter (a whole desktop page as a browser loads it). 1 GB is counted here as 1,000,000 KB. Your targets will differ from the median: online shops with large product images weigh more, API endpoints far less.

The comparisons’ “How much do I need?” box does this arithmetic with your own numbers and sends the result straight into the ranking.

What makes a scrape heavier than it looks

  • Rendering in a browser. A headless browser loads images, fonts, scripts and the calls those scripts make, often dozens of times the size of the HTML. Block images and fonts in the browser when you only need the text.
  • Retries and blocks. A blocked request still transfers a page, often a challenge page, and the retry transfers another. A target that blocks half your requests can nearly double the traffic.
  • Redirects and sessions. Each redirect is another request; a scraper that starts a fresh session for every page downloads the same cookies and scripts again.
  • Compression. Asking for compressed responses (most HTTP clients do by default) cuts HTML traffic several times over; check that yours does.

How to measure your own job

Run a sample of a few hundred real requests through the proxy you plan to use, then read the usage in the provider’s dashboard and divide by the number of requests. That gives your own size per page, retries included, as the provider meters it. Multiply by the pages you plan a month.

From gigabytes to a price

Type the result into the residential or mobile comparison. Every plan that covers it is ranked by what you pay at checkout, with what expires and what renews. If your scraper runs every month, set the amount to “GB a month” so packs that last are spread over the months they cover. What does a proxy cost for one scraper for a month? walks through the rest of the choice.

Price it for your amount

Checkout totals, every billing type competing. Change the amount to recalculate.

Geonode

$34.65

Residential monthly subscription · Monthly

50 GB needed50 GB included
Rolls over while subscribed · 0 GB unused

Checked Nov 17, 2026 · $0.69 per GB used

Buy at Geonode →

GB and IP packages use separate amounts and billing units.

On this page