The world needs more data. Now it can afford it.

AI made it possible to use far more of the web than anyone collected before, while proxy and scraping prices stayed where they were. We re-engineered our proxy network and Scraping API from the hardware up to change that.

AI is going into every company.

A language model can read, sort and summarize more data in an afternoon than an analyst could in a month. Every company can now analyze far more data, so it needs to collect far more of it. AI product companies feed agents, AI search and training pipelines. E-commerce teams compare more prices, SEO teams track more results and research teams test more signals.

The price per gigabyte did not keep up.

Most of the industry still prices web data for the volumes teams collected before AI. At those rates, a team that needs ten times the data pays ten times the bill.

What we changed.

We obsessed over the low-level details to build new pricing for a new era. We have built proxy networks since 2015, so we knew which costs could come down: cloud rates for heavy compute, retail hosting rates and a slow request path. We own our residential network, and heavy compute runs on bare-metal servers we own and operate. We move petabytes of traffic a month, which earns us volume rates on infrastructure and hosting. And we wrote the request path in Go, so a single gateway server handles thousands of requests a second.

We kept our margins where they were and made lowering our costs a goal for the whole company, so the savings could go to the teams collecting data in the AI era. For larger volumes, our sales team offers further discounts depending on commitment and use case.

The world needs more data. Now it can afford it.