Random Number Generator

They come from crypto.getRandomValues, the browser's cryptographically secure generator — the same source that creates encryption keys, seeded by the operating system from hardware events. That is a categorically stronger source than Math.random, which is designed for speed, not unpredictability, and should never decide anything with stakes.

The obvious way to map a random 32-bit value onto a range — value % range — is subtly unfair whenever 2³² does not divide evenly by the range: the low numbers come up slightly more often. This tool uses rejection sampling instead: draws that would fold unevenly are discarded and redrawn, making every value in your range exactly equally likely. It is the difference between a tool that looks fair and one that is.

It draws without replacement, like pulling numbered balls from a bag: each value can appear at most once. That is the mode for raffles, prize draws and assigning people to positions. Asking for more unique numbers than the range holds — ten distinct values between 1 and 5 — is refused rather than fudged.

Mechanically, yes — the draw is cryptographically random and without-replacement mode guarantees distinct winners. Practical fairness is about process, not just randomness: fix the participant list and announce the rules before drawing, map numbers to entries publicly, and draw once. The tool cannot prove to your audience that you did not redraw — a screen recording of a single draw is the usual answer.

Above that, a browser textarea stops being a sensible container and the use case is really data generation — a job for a script. The range itself can span billions; only the count of results per draw is capped.

No — generation happens entirely in your browser, which is also why the page keeps working offline. Nobody else sees your numbers, and no server log could reconstruct a draw.