Random Number Generator — Free Online Integer & Decimal Generator

What is Random Number?

Random number generation looks simple but has real pitfalls. Naive `Math.floor(Math.random() * range)` is fine for a coin flip but suffers modulo bias when range doesn't evenly divide the underlying 2^32 output space — some numbers become slightly more likely than others. The crypto-strong mode of this tool avoids that with rejection sampling: it discards raw output that falls in the 'imbalanced tail' and resamples until the value is in a range that divides cleanly. For casual use (dice rolls, game randomness, sample data), the default Math.random path is fast enough. For anything security-adjacent (unique tokens, unbiased research samples), flip on the strong mode.

How to generate random numbers online

1. 1Set the min and max — inclusive on both ends.
2. 2Set how many to generate.
3. 3Choose integer or decimal (with configurable decimal places).
4. 4Toggle 'unique values only' for lottery-style draws, or 'cryptographically strong' for unbiased randomness.

Use Cases

Key Concepts

Modulo bias

The distribution-skewing that happens when you take `randomInt() % N` where N doesn't divide the range evenly. Small skews are invisible in games but destroy statistical validity in security or research.

Rejection sampling

Discard raw random samples that fall in the biased tail and resample. Guarantees uniform distribution at the cost of occasional re-rolls. Standard fix for modulo bias.

Tips & Best Practices

• ▸Math.random() uses a linear congruential PRNG in most engines — perfectly fine for games, not for cryptography. crypto.getRandomValues uses OS-level entropy, suitable for tokens and keys.
• ▸The 'unique values only' path can't produce more values than the range allows. If you ask for 1000 unique in [1, 500], the tool will refuse and explain.
• ▸Modulo bias is real: `crypto.getRandomValues([0])[0] % 6` slightly favors 0-3 over 4-5 because 2^32 doesn't divide by 6. The crypto-strong mode uses rejection sampling to fix this.
• ▸For reproducible sequences (unit tests, benchmarks), seed a PRNG in code — this tool always uses fresh randomness.