RunToolRunRandom Number Generator
Generate random numbers.
How to Generate Random Numbers
Set your range
Enter the minimum and maximum values — 1 to 100, 1 to 6, anything.
Choose how many numbers
Generate one number or a whole batch at once.
Generate and copy
Numbers appear instantly and can be copied with one click.
Where the randomness comes from
Ordinary programming random-number functions are built to be fast and evenly distributed, not unpredictable. Given enough output, their internal state can be reconstructed and every subsequent value predicted — which is fine for a dice roll and not fine for a prize draw where someone has an incentive to guess.
This generator uses the browser's cryptographically secure source instead, seeded from operating-system entropy and designed so that past output reveals nothing about future output. The difference is invisible in the result: two sequences can look identical and differ completely in how hard they are to predict.
Random does not look random
People reject genuinely random sequences as broken because they contain patterns. In twenty coin flips, a run of four or five in a row is likely rather than surprising. Among ten numbers drawn from one to a hundred, a repeat or two adjacent values is common.
This is why humans asked to 'be random' produce sequences that are far too even, avoiding repeats and clusters. A generator that never repeated would be the suspicious one — clustering is a property of randomness, not evidence against it.
Ranges, inclusivity and bias
Whether a range includes its endpoints changes the odds, particularly on small ranges: one to six inclusive is six outcomes, and getting that wrong silently removes a face from your die. The maximum here can occur.
There is also a subtler issue that most implementations get wrong. Mapping a random value onto a range with a remainder makes some outcomes slightly more likely than others whenever the range does not divide the pool evenly — modulo bias. This generator avoids it by discarding values that fall in the uneven tail and drawing again, so every number in the range is exactly as likely as every other.
Independence, and the gambler's fallacy
Each draw knows nothing about the ones before it. A number that has appeared three times running is exactly as likely to appear again as any other, and a number that has not appeared for a long time is not 'due'. This is the most persistent misunderstanding about randomness and it costs people money regularly.
If you need draws without repetition — picking five different winners from a list — turn duplicates off. That switches to drawing without replacement, which is a genuinely different operation: repeatedly generating numbers and throwing away the ones already seen gets slower and slower as the draw fills up, and drawing ten numbers from a range of ten that way takes about twenty-nine attempts on average.
Fairness in practice
For a public draw, the arithmetic matters less than the process. Announcing the method and the range before drawing, and generating in front of witnesses or on a recording, is what makes a result credible. A perfect generator used privately convinces nobody.
Generation happens in your browser with no server involved, which means no one — including this site — logs what you generated or when. That is a privacy property, not a fairness guarantee.