Random Number Generator

The Random Number Generator produces random integers or decimals within any range you specify. Set min, max, count, and whether you want integers or decimals, and the tool returns your requested random values. Use for simulations, test data, raffles, dice-rolling, lottery-number picking, sampling from a dataset, Monte Carlo methods, or any task needing unpredictable numbers.

Output is cryptographically strong by default (via crypto.getRandomValues) with an option to use Math.random for non-security contexts where speed slightly beats strength. Bulk mode generates up to 10,000 numbers at once with optional uniqueness (great for lottery-style picks without replacement). Seeded mode gives reproducible sequences for debugging and testing. Distribution options include uniform (default, every value equally likely), normal (Gaussian bell curve), and weighted (specify probabilities per value). Everything runs in your browser — seeds and generated numbers never leave your machine.

Random integer in range: use (crypto random as 0-1 float) × (max - min + 1) + min, then floor. This gives a uniform distribution where every integer in the range is equally likely.

To avoid modulo bias with byte-level randomness: rejection sampling. Generate a random 32-bit integer; if it is outside the acceptable range, reject and try again. This preserves exact uniformity at the cost of occasional retries.

Random decimal in range: (crypto random as 0-1 float) × (max - min) + min. Provides roughly 52 bits of floating-point precision (IEEE 754 double).

Uniqueness (sampling without replacement): use Fisher-Yates shuffle on a range array and take the first N values. Or use a Set-based approach that rejects duplicates during generation.

Seeded random: for reproducibility, use a seeded pseudo-random generator like mulberry32 or xoshiro. Seeded random is not cryptographically secure but is reproducible, useful for testing and debugging.

Normal distribution: Box-Muller transform takes two uniform random samples and produces two normally distributed samples with mean 0 and standard deviation 1. Shift and scale to any mean/stddev.

Weighted random: specify weights for each option. Sum weights, pick a random value in [0, total). Iterate options, subtracting weight from random value until it goes negative — that is the selected option. O(n) per selection but handles any weight distribution.

Bulk uniqueness: when generating many unique numbers, memory usage scales with count. For millions of numbers in a huge range, this is fine (you just need indices). For unique numbers in a small range, you can\u2019t have more than the range size.

Performance: generating 10,000 cryptographic random integers takes around 50-100 ms. Math.random-based is about 10x faster but not cryptographically strong.

Compared to programming language built-ins, this tool is the interactive version with uniqueness, distributions, and entropy already handled. For automation, use Python\u2019s secrets or JavaScript\u2019s crypto API in code.

Compared to online dice or lottery tools, this one is more flexible (any range, any count) and uses cryptographic randomness by default. Simple dice tools fix the range; this one doesn\u2019t.

Compared to spreadsheet RAND or RANDBETWEEN, this tool handles cryptographic strength, uniqueness, and larger batches. Spreadsheets are better for mixing random numbers with other formulas; this tool is for pure random generation.