Math Expression Calculator

Why This Calculator Uses a Custom Parser

The shortest way to compute a string like "2 * (3 + 4) ^ 2" is a single line of code: eval(expression). But this one line is a notorious security hole. That's because eval doesn't just calculate expressions—it executes any string as JavaScript code.

The moment you accept user input, your calculator becomes an "arbitrary code executor." In a browser, a user could run code to scrape cookies and login tokens and send them to an external server (XSS). If a calculator API on a server were built with eval or new Function, it would be even more catastrophic—it could read files, execute commands, and completely take over the server. A simple calculator feature would create a remote code execution vulnerability.

That's why this tool uses neither eval nor new Function. It breaks the expression down into tokens (with a tokenizer), organizes the order of operations using the Shunting-yard algorithm, and then calculates the result using only allowed functions and operators. If it encounters an undefined name, it throws an error instead of executing it—which is why entering alert(1) is rejected with an "Unknown name" error.

Try Entering 0.1 + 0.2

Enter 0.1 + 0.2 into this calculator, and you'll get 0.3. But if you do the same calculation in your browser's console, you'll see 0.30000000000000004. Is the calculator wrong? No—both computed the same value. This tool just polishes the final number before displaying it.

The cause is how computers store decimal numbers in binary. The decimal 0.1 becomes a repeating fraction in binary, so it has to be truncated at some point. This tiny error becomes apparent during addition. This isn't just a JavaScript problem; it happens in almost every language that uses floating-point numbers. This tool just hides the noise by formatting the result to 12 significant digits, but the internal calculation error is still there.

So, How Should You Calculate Money?

There is one main principle: handle money as integers, not decimals. Perform calculations using the smallest currency unit (like cents), and only apply explicit rounding or truncation rules when division is necessary.

Value-added tax (VAT) is a classic example. If you need to find the pre-tax price from a total of $290.00 with 10% tax, you get 290 / 1.1 = 263.6363... Whether to round this up or truncate it isn't a math problem—it's a business and tax regulation problem. A calculator can't decide that for you; the rule must be centralized in your code.

• In your database, use DECIMAL or NUMERIC columns instead of FLOAT or DOUBLE. Using FLOAT for money is a common but dangerous mistake.
• When handling money in JavaScript, convert amounts to integers (e.g., cents) or use a dedicated library.
• Don't accumulate the raw results of division. Finalize the rounding rule at each step before proceeding to the next calculation.
• This calculator is great for sanity checks and estimations. Don't run your actual accounting logic on it.

Five Common Syntax Gotchas

The error message will tell you the approximate character position of the problem, but the issues below can be tricky to diagnose if you don't know what to look for.

Large Number Limits and History

If a result exceeds 9,007,199,254,740,991 (2^53 - 1), a warning appears: 'Value exceeds 2^53, so the last digit is an approximation'. This is because JavaScript's standard number format cannot accurately represent integers larger than this. Calculations like 2^100 will produce a value, but it won't be an exact integer. For arbitrary-precision arithmetic, use a language like Python or an environment that supports BigInt.

Pressing Enter after entering an expression saves it to the Calculation History. It stores up to 10 entries. Clicking an item in the history reloads that expression into the input box, which is handy for repeating similar calculations with different values.

• History and your last entry are saved only in this browser. They are not sent to the server and will not be visible on other devices.
• If you've calculated sensitive numbers on a public computer, be sure to click "Clear History" when you're done.
• Undefined calculations like division by zero, the square root of a negative number, or tan(90°) are caught as errors instead of producing strange values like Infinity or NaN.
• The chip buttons (pi, sqrt(, round(, etc.) insert text directly at your cursor's position. You don't need to memorize the function names.