The company behind
the API.
Maskbreak is built by Sentinel Edge Networks LTD, a company registered in England & Wales. We build one product: a real-time fraud detection API that tells you whether a visitor is hiding behind a VPN, proxy or Tor — naming the service when known — or using a fake browser or bot, in under 150 milliseconds — and we publish the research behind it.
Who runs it
Maskbreak is founded and run by Kaspar Tomson. He runs the detection pipeline and the antidetect harness: he buys and operates the same antidetect browsers, proxy services and automation frameworks that fraud operators use, points them at the detector, and publishes what fired and what did not. Articles on the research blog credited to him are written by him. Find him on LinkedIn and GitHub, or through the contact page.
What we build
Most fraud tooling was designed for an older internet: datacenter IPs, obvious bots, known-bad blocklists. The abuse we focus on looks different — commercial fake (“antidetect”) browsers such as Kameleo, GoLogin, Multilogin, AdsPower, and Dolphin{anty}; residential proxy networks that route attacks through clean home IPs; and AI-driven automation that mimics human behavior. Maskbreak evaluates the network layer and the device layer together in a single API call, so the operator running five hundred “different” accounts from one laptop stops looking like five hundred different users.
The product is documented in the API reference, free — see the free page (1,000 requests an hour, every signal, no card) — and you can try it without signing up: the homepage scanner checks your own connection live, and the free IP lookup tool checks any address against Tor and cloud-server ranges.
How the research desk works
The Maskbreak research blog exists because we test things. We buy and run the same antidetect browsers, proxy services, and automation frameworks that fraud operators use, point them at our own detector, and publish what fired and what did not — including the live tests where we show the scanner output signal by signal. When a claim in an article is measurable, we prefer to measure it.
Articles credited to a named person are written by that person. Maskbreak Research is the byline for work produced and reviewed by our internal research desk rather than a single author. Product and pricing claims are checked against the vendor's official documentation on the article's stated update date, and comparison pages link the sources that support material claims.
Lab articles describe the tested version, setup, sample, and observed result so a test is not presented as a universal detection rate. Comparison content also names its limits: where a competitor's public documentation shows no coverage for something, we say “we have not seen documented detection,” not “they cannot do it.” We disclose that we are biased toward our own product and do not publish an accuracy percentage without a reproducible benchmark.
See something wrong or out of date? Send the URL and correction through our contact page. We verify the report, update the page and its modified date when the substance changes, and leave claims unchanged when the evidence does not support the correction.