Bandwidth Cost Calculator

requests per day × days per cycle × average payload size in KB, converted to GB, then multiplied by (1 + overhead percentage). The result is the total transferred volume for the billing cycle. Multiply by your per-GB price to get the USD total.

10% for pure JSON APIs, 15–20% for HTML scraping with normal retries, 25–30% for headless-browser workloads with asset loading, up to 40% for aggressive retry policies. When in doubt, start with 20% and adjust after your first billing cycle.

Yes — most residential and mobile SOCKS5 providers meter total bytes transferred through the tunnel, in both directions. Request bodies are usually small (headers only) but count toward the total. The calculator's average-response-size input covers the outbound side implicitly for typical HTTP flows.

Load 10–20 target URLs in Chrome DevTools and note the 'transferred' column in the Network tab. Take the median (not the mean — outliers skew averages). For JSON APIs this is usually 5–50 KB; static HTML 30–200 KB; JS-heavy SPA pages 500 KB–3 MB.

Residential and mobile pools (Bright Data, Oxylabs, Smartproxy, IPRoyal, SOAX, Nimble) bill per GB — usually $2.50–15/GB depending on plan size. Datacenter and ISP-static providers (Rayobyte, Proxy-Cheap, PrivateProxy) bill per port or per IP with unlimited bandwidth.

The break-even is around 80–120 GB per month. Above that, unlimited datacenter plans ($75–150/mo) or ISP-static plans ($1.50–3/IP/mo) usually beat residential on raw cost — as long as your target does not fingerprint datacenter ASNs.

Yes — that is what the overhead percentage covers. A 20% overhead assumes ~20% of requests retry once. If your target has a high block rate and you retry 3× per URL, raise the overhead to 40–60% and re-check.

Yes. Every input syncs to the page URL. Click Share to copy a link that reproduces the exact scenario. Useful for comparing quotes from multiple providers or documenting a budget review.

Enter the daily request count during your active window (not spread across 24 hours). Overhead percentage is unchanged. If bursts hit provider concurrency ceilings, add a 5–10% overhead premium for retried requests that hit throttle limits mid-burst.

Only if your scraper actually loads them. Headless-browser scrapers (Playwright, Puppeteer with full-page rendering) do — include CSS, JS and image bytes in your average size. HTTP-only scrapers (requests, cURL) skip subresources — use the raw HTML transfer size from DevTools' Network tab instead.

Compute each pool separately with its own per-GB price, then sum the totals. Residential at $8/GB × 30 GB + datacenter at unlimited $99 flat = $339/mo total. The Share URL lets you save each scenario separately for finance review.

Run a 1-day pilot scraping 5–10% of your target volume, measure actual GB transferred at the proxy dashboard, then multiply. Feed that measured GB/day into the calculator with a 30-day cycle and a realistic overhead. Real measurements beat theoretical page-size estimates by 20–40% in either direction.

Sticky-session fees ($0.20–0.50 per session on some providers), premium geo surcharges (residential in tier-3 countries costs 2–3× the base rate), and captcha-solving credits if the target requires it. Add these to the per-GB total in your final finance sheet even though the calculator focuses on pure bandwidth.

Run for one full billing cycle, note actual GB from the provider invoice, and reverse-engineer the overhead: overhead = (invoice GB ÷ (requests × avg size in GB)) − 1. Feed that measured overhead back into the calculator for future forecasts — you'll get an accuracy inside 5%.