A lens, not a topic
Most tools listed here have a primary home somewhere else. The unit converters also sit under Converters; the dice, coins and draws also sit under Randomizers. They appear here because at some point you think of the task as arithmetic rather than as a conversion or a game, and this is where that instinct lands you.
Working with quantities — length, area and angle, including the degrees-to-radians step that trips up anyone calling a trigonometry function for the first time.
Working with how numbers are written — binary, octal, decimal and hexadecimal, and Roman numerals in both directions.
Working with numbers as objects — the prime number generator, which lists primes in a range and shows the sieve doing it.
Producing numbers — a range you define, dice notation, a coin, lottery numbers, bingo calls, and a verifier for a draw somebody else ran.
Exact answers and approximate ones
Worth knowing which kind you are getting, because the tools do not look different.
Exact. Number system conversion, Roman numerals and prime generation are exact — a number has one binary spelling and 97 either is prime or is not, and no rounding is involved anywhere. Some unit conversions are exact too, by definition rather than by measurement: one inch is defined as precisely 25.4 mm, one mile as 1609.344 m. Those relationships are not measured, they are agreed.
Approximate. Most other conversions are decimal expansions cut off somewhere. Converting to a unit and back does not always return exactly what you started with, and chaining conversions compounds it — metres to feet to inches to centimetres accumulates four roundings. If it matters, convert once, directly, from the original figure rather than from a converted one.
Degrees and radians deserve a warning of their own, because the ratio is irrational. π radians is exactly 180°, but π itself never terminates, so every decimal radian value you see is truncated. This is why 90° comes back as 1.5707963… and why comparing two radian values for exact equality in code is a bug waiting to happen.
What is not here
No statistics, no equation solver, no matrix arithmetic, no graphing. This category is small operations on individual numbers rather than a maths package. For anything symbolic, WolframAlpha or a computer algebra system is the right tool and this is not trying to compete with them.