Scrapoxy

Scrapoxy is not a traditional proxy seller — it was an open-source proxy aggregator and orchestration layer that sat between your scraper and the dozens of upstream proxy providers you were already paying for. Built by French engineer Fabien Vauchelles and first demoed at HumanTalks Paris in September 2015, Scrapoxy gave scraping teams one HTTP endpoint to hit, then transparently rotated, sticky-sessioned, and load-balanced traffic across every connected provider behind the scenes. Important update for buyers: Scrapoxy was officially discontinued on February 6, 2026 after more than eleven years, so this entry is more of a historical record and a guide on how to evaluate the surviving alternatives.

On the network side, Scrapoxy never owned IPs of its own. Instead it shipped 50+ connectors to commercial proxy vendors (Bright Data, Decodo, IPRoyal, NetNut, Rayobyte, Live Proxies, Proxy-Seller, ProxyScrape and others), plus connectors to cloud providers like AWS, GCP, Azure, DigitalOcean, OVH, Scaleway and Vultr so you could spin up datacenter proxies on demand. Through the platform it tracked roughly 19 million IPs across its lifetime and handled around 115 billion requests for 1,742 known deployments — credible numbers for what was effectively a one-engineer project. Geographic coverage and city-level targeting were inherited from whichever upstream vendor you plugged in, not provided directly by Scrapoxy.

Pricing was the unusual part. The core software was free and self-hosted via Docker, which is why it became a staple in scraping stacks at agencies and data teams who already had a residential or ISP plan elsewhere and just needed orchestration. A paid enterprise support tier and dedicated hosted infrastructure were introduced later, but uptake stayed thin — the founder publicly cited "a market that structurally resists paying for this kind of tooling" as the reason for shutting down. Existing enterprise subscribers keep their hosted backend until contract end; no new subscriptions are being sold, public Docker images have been removed, and forking is explicitly prohibited under the commercial license.

The standout features were the ones that made Scrapoxy genuinely loved by web-scraping engineers: a vendor-agnostic dashboard for monitoring every IP source in one place, automatic blocked-proxy detection that pulled dead IPs out of rotation, sticky-session control per request, freeze/quarantine logic for IPs that got soft-banned, and native integration with Playwright, Puppeteer, Scrapy, and any HTTP client that speaks a standard proxy URL. It also supported HTTP, HTTPS and SOCKS connectors and let you write a custom connector in TypeScript for any vendor it didn't ship with. For teams running anti-bot bypass against Cloudflare, DataDome, or PerimeterX, the value was being able to A/B residential pools from three vendors against the same target and route around whichever one was burning JA3 signatures that week.

The ideal customer for Scrapoxy was always a mid-sized scraping team or agency that bought proxies from two or more vendors and wanted one control plane — not a small operator buying their first GB of residential traffic. If you are that buyer in 2026, your realistic alternatives are commercial proxy management layers from Decodo, Bright Data Unblocker, Oxylabs Web Unblocker, or scraping APIs like ScrapingBee, Zyte, and Scrape.do, plus self-hosted open-source orchestrators like Crawlee proxy rotation or Proxidize for mobile. Scrapoxy's niche — open-source, vendor-agnostic, self-hosted — currently has no true successor, and the founder has confirmed there will be no community takeover.

Limitations to know up front: because the project is now end-of-life, you should not build new infrastructure on Scrapoxy. The shared backend services for free users (GeoIP resolution, public proxy health checks) are gone, Docker images have been pulled from public registries, documentation is being archived, and the source is proprietary so a community fork is not legally on the table. Even before shutdown, Scrapoxy assumed you had real scraping engineering capacity in-house — it was never a one-click solution, and the configuration surface for connectors, freezers, and rotators was substantial. Treat any forum thread suggesting Scrapoxy as a current recommendation with caution and verify the end-of-life timeline at scrapoxy.io before evaluating it for new work.