Send a URL, get HTML or plugin JSON back. JoyProxy handles egress, retries, and optional JS rendering. You do not buy proxy GB and you do not run Chrome.
Pay credits only when the fetch succeeds. Keep rotating proxies for a later day — they are for teams that already operate a crawler.
POST /v1/fetch from your service. No SOCKS config.
Paste a URL in API Center, download HTML, move on.
Wire API Center output into sheets, n8n, or a warehouse.
Need rendered DOM or Amazon/Google/YouTube plugin JSON.
If you must attach a partner-whitelisted IP, or you already have a tuned Playwright farm, buy proxy IP instead.
Patch Tuesday, RAM, and CAPTCHA vendors become a second product team.
Target rules change weekly. You want that burden on the API, not your repo.
Proxy GB still bills when the page is empty. API credits do not.
Missing geoCode returns the wrong storefront with a 200.
Multi-step login needs sessionId — do not fake it with rotating GB.
Credits per success are easier to forecast than GB plus retries.
Recommended: Web Scraping API. Alternative: rotating residential only if you already run Scrapy/Playwright and just need IPs.
| Product | Billing | Use it when… |
|---|---|---|
| Web Scraping API Recommended | Credits · pay per success | Recommended — /v1/fetch, render, plugins, pay on success. |
| Residential · Rotating | Pay per GB | Alternative — you already have a crawler and only need egress. |
Prove one URL before writing code.
Do not force the API. Buy rotating residential instead.
Not for /v1/fetch. Buy proxy GB only if you also run your own clients.
Yes — separate tokens for API vs extract.
Amazon, Google Search, YouTube return JSON. Everything else is raw HTML.
API egress is managed. For a fixed allowlist IP, buy static residential instead.
API Center and OpenAPI Center — not the proxy Usage tab.
Credits apply only on success. Skip proxy packs until you know you need them.