Graphing Calculator
Plot up to three functions at once, set the x-window, and read roots, extrema and undefined points straight from the graph. Handles powers, sin, log and exp.
Graphing Calculator
x^2 - 4
- Window minimum
- -3.9994
- Window maximum
- 96
- Approximate roots
- -2, 2
- Undefined points
- 0
sin(x)
- Window minimum
- -1
- Window maximum
- 1
- Approximate roots
- -9.425, -6.283, -3.142, 0 (+3 more)
- Undefined points
- 0
| x | x^2 - 4 | sin(x) |
|---|---|---|
| -10 | 96 | 0.544 |
| -6.667 | 40.1109 | -0.3508 |
| -3.333 | 7.1111 | 0.1906 |
| 0 | -3.9994 | 0.0251 |
| 3.333 | 7.1111 | -0.1906 |
| 6.667 | 40.7792 | 0.3973 |
| 10 | 96 | -0.544 |
In 1985 Casio shipped the fx-7000G, the first handheld graphing calculator, and within a decade the device had reshaped how mathematics was taught: instead of plotting a handful of points by hand on squared paper, a student could see the whole shape of a function in one second. The same shift has now moved to the browser. A graphing calculator takes an expression such as x^2 - 4, evaluates it at hundreds of x-values inside a window you choose, and draws the resulting curve so that roots, peaks, troughs and breaks become visible at a glance. [desmos]
The reason a graph is so useful is that algebra and geometry answer different questions. An equation like x^2 - 4 = 0 tells you that solutions exist; its graph tells you that the parabola crosses the horizontal axis exactly twice, at x = -2 and x = 2, and that between those crossings the function dips to its lowest point at (0, -4). One picture replaces a page of sign analysis. [openstax-functions]
This calculator plots up to three functions simultaneously. You type the expressions, choose the horizontal window from x-min to x-max, and the tool samples each function, draws the curves with a shared coordinate system, and reports a compact summary: the smallest and largest values reached inside the window, the approximate roots, and how many sampled points returned no finite value. It is deliberately a visualisation and inspection tool, not a computer algebra system, and the Limitations section is explicit about where that line sits.
Students use it to check homework and to build intuition about families of functions. Engineers and analysts use it to sanity-check a model before trusting a spreadsheet, because a plot exposes a sign error that a table of numbers can hide. Anyone who needs to answer "what does this expression actually look like" can use it without installing anything, and the companion Derivative Calculator picks up where the picture leaves off, telling you the slope of the curve at a chosen point.
A function pairs every input in its domain with exactly one output. The graph is the set of all those pairs drawn as points in the plane, with the input on the horizontal axis and the output on the vertical one. That single sentence carries two constraints worth spelling out, because most plotting mistakes come from ignoring one of them. [mathworld-graph]
The first is the domain: the set of inputs the expression can accept. sqrt(x) has no real output for negative inputs, ln(x) only exists above zero, and 1/x is undefined exactly at zero. A plotting tool cannot show what does not exist, so it simply skips those samples, and the curve visibly stops or breaks. The second is the range: the set of outputs actually reached. When a summary reports a minimum of -4 and a maximum of 96 inside a window, that is the range restricted to that window, not the range of the function over all real numbers.
The vertical line test follows from the definition: if a vertical line crosses the drawing twice, the relation is not a function of x. A circle fails the test, which is why circles are drawn as two separate branches, an upper and a lower half, in every graphing tool that plots explicit functions.
It also helps to separate three things that beginners often merge. The expression is a rule. The graph is a picture of the rule over a chosen interval. The window is a viewing decision, not a property of the function. Change the window and the picture changes dramatically while the function stays exactly the same: a parabola looks like a narrow spike in one window and like a gentle bowl in another.
The layout is intentionally minimal: three expression fields, two window fields, and a plot that updates as you type.
- Type an expression in the first field, using
xas the variable. Standard operators and functions are available, and multiplication can be left implicit:2x,3(x + 1)andx(x - 1)are all accepted. - Optionally add a second and third expression. Each one gets its own colour in the plot, its own summary card, and its own column in the values table.
- Set the horizontal window with the two number fields, or pick a preset such as minus ten to ten, minus one to one, zero to two pi, or zero to one hundred.
- Use the zoom buttons to halve or double the width of the window around its centre without retyping the limits.
- Read the values table below the plot for seven evenly spaced points, which is often faster than hovering over the curve.
Example 1 - A parabola. Plot x^2 - 4 in the default window from -10 to 10. The summary card reports a minimum of -3.9994, which is the sampled approximation of the true vertex at -4, and approximate roots at -2 and 2, which matches the factored form (x - 2)(x + 2). The table shows 96 at both ends, because the window ends at plus or minus ten.
Example 2 - A periodic function. Switch the window to 0 to 2 pi and plot sin(x). The curve rises to 1, crosses zero at pi, falls to -1, and returns to zero at 2 pi. The root readout shows 3.142, which is pi to three decimals and a fair demonstration that the sampling grid is fine enough to land on a crossing with visible precision. The Trigonometry Calculator covers the same function a value at a time, which is the right tool when you need one ratio rather than a whole curve.
Example 3 - A discontinuity. Plot 1/x in a window from -5 to 5. The curve appears to jump from the bottom of the frame to the top at the origin, but the summary will usually report zero undefined samples, and that mismatch is instructive: with four hundred points across ten units the grid steps by about 0.025, so it almost never lands exactly on x = 0. The tool can only report what it sampled. Treat a vertical asymptote as a visual observation, not as a measurement, and check the offending x-value by hand.
Example 4 - Comparing families. Put x^2, x^3 and 2^x in the three fields at once. Inside a small window the cubic grows fastest on the positive side; widen the window to 0 to 100 and the exponential overtakes everything. Seeing that crossover is far more instructive than memorising "exponentials eventually win".
Three features do most of the interpretive work in a graph, and the summary card addresses each of them with the same honesty: it reports what the sampled grid can see, not what a symbolic solver would prove.
A root is an input where the output is zero. The tool samples the window at evenly spaced points and flags every adjacent pair whose outputs have opposite signs, then estimates the crossing by linear interpolation between those two samples. [mathworld-graph] If the outputs at two neighbouring samples have opposite signs, the straight line joining them crosses the axis at
That estimate is accurate to the spacing of the grid for well-behaved curves, and it degrades where the function is nearly vertical. It also misses roots where the curve touches the axis without crossing it, such as x^2 at the origin, because the sign never changes. This tool will report no root there, and the graph itself is the correction. The opposite error is just as real: a sign change produced by a jump such as 1/x at the origin is reported as a root even though the function has no value at that point, so a root that sits next to a vertical asymptote deserves a manual check.
An extremum is a local maximum or minimum. The summary reports the largest and smallest values found in the window, which is a sampled approximation of the global extremes over that interval. When the true extremum sits between two samples, the reported value is slightly smaller or larger than the exact one. For a grid of four hundred points the error is invisible on screen, but it is not zero, and anyone who needs the exact vertex should complete the square or use the Quadratic Equation Calculator. At a minimum the curve is momentarily flat, which is why the Slope Calculator is the natural companion: a slope of zero is the algebraic signature of the turning point the plot shows.
An asymptote is a line the curve approaches without meeting. Vertical asymptotes show up as undefined samples and as branches that shoot off the top of the frame; horizontal asymptotes show up as a curve that flattens. Neither is announced by the tool, because detecting them reliably requires limits rather than samples. The visual signal is strong, though: a curve that keeps leaving the frame on both sides of one x-value has a vertical asymptote there.
The density of the samples determines what the numbers mean. With N points across a window of width W, the spacing is
and every root estimate carries that granularity as its uncertainty. Widening the window without adding samples coarsens the grid, which is why zooming in usually improves both the picture and the numbers.
Four hundred samples across a window give a step of 0.05 for the default range from -10 to 10. For the shapes most people plot, that is plenty: a parabola, a cubic, a logarithm or a sine with an ordinary period all change slowly relative to that step, so the polyline drawn between consecutive samples is visually indistinguishable from the true curve.
The failure mode is aliasing. A function that oscillates faster than the sampling grid can follow will be drawn as a misleading smooth wave, because the samples land at similar phases each time. Plot sin(50x) in the default window: its period is about 0.126, the grid step is 0.05, so fewer than three samples fall inside each period, and the plot shows a slow ripple instead of a dense oscillation. The cure is the same in every numerical context: narrow the window so the interesting variation gets more samples per period. [dlmf]
The same limitation explains sharp corners. A function with a vertical tangent, an absolute value kink, or a step discontinuity is drawn as a straight segment across the feature, because the tool only knows the values at the sampled points. Zooming in repeatedly is the honest way to inspect such a feature; the drawing never becomes a proof, only a better approximation.
Finally, the vertical scale is chosen automatically from the sampled values, with extreme outliers trimmed so that one asymptote does not flatten everything else into a horizontal line. That is why a curve near an asymptote may appear to stop before leaving the frame entirely: the visible band is a viewing choice, not a statement about the function.
The parser accepts ordinary mathematical notation. It does not execute JavaScript, and it does not evaluate arbitrary code, which matters because expressions are typed by visitors rather than by developers.
| Variable | x | the horizontal axis value |
| Constants | pi, e | 3.14159 and 2.71828 approximately |
| Powers | x^2, x^0.5 | raising to a power, evaluated right to left |
| Implicit product | 2x, 3(x+1) | multiplication written without a symbol |
| Basic functions | sqrt, cbrt, abs, exp | square root, cube root, absolute value, e raised to x |
| Logarithms | ln(x), log(x), log2(x) | natural, base ten, base two |
| Trigonometry | sin(x), cos(x), tan(x) | radians, with inverse and hyperbolic variants |
| Rounding | floor(x), ceil(x), round(x) | step and integer functions |
Implicit multiplication is the feature people notice first, because it lets an expression be typed the way it is written on paper. The parser inserts the multiplication when a completed operand is followed directly by another operand or by an opening parenthesis. [openstax-functions] Precedence follows the usual convention, so powers bind tighter than multiplication, which binds tighter than addition, and a leading minus sign applies to the whole power:
Very large or very small coefficients are easier to type in scientific notation, and the Scientific Notation Calculator converts between the two forms when a plot has to be read against a textbook value. For the arithmetic of a single power, the Exponent Calculator returns the exact result.
Anything the parser cannot read produces a message naming the problem, such as an incomplete expression, an unknown function, or unbalanced parentheses, and the plot waits for a valid expression rather than displaying a guess.
The window is the single decision with the largest effect on what you learn, and there are three reliable strategies.
Strategy one: bracket the feature you care about. If you are looking for roots, choose limits where the function changes sign on both sides, then let the interpolation do the rest. If you are looking for a vertex, centre the window on your estimate and narrow it until the minimum becomes visually obvious.
Strategy two: start wide, then zoom. A too-narrow window can hide the entire feature you were looking for: the curve may enter and leave the frame between two x-values you never sampled. A wide first look costs nothing and prevents that mistake. The zoom buttons keep the centre fixed, so repeated clicks converge on the interesting region without retyping numbers.
Strategy three: use a table when the picture is ambiguous. A flat-looking region in a plot may be a genuine plateau or a subtly increasing line. The values table resolves the question immediately, because it reports seven numbers with enough digits to separate them.
The values of the parabola used throughout this page, computed at integer inputs from -3 to 3, show why a table and a picture complement each other. The table gives exact numbers; the chart gives the shape that makes those numbers meaningful.
| -3 | 5 |
| -2 | 0 |
| -1 | -3 |
| 0 | -4 |
| 1 | -3 |
| 2 | 0 |
| 3 | 5 |
Reading the two together is the habit worth building. The table proves the symmetry; the chart shows that the symmetry comes from a smooth bowl rather than from a coincidence of the seven chosen inputs.
- Check the summary before trusting the picture. A curve that looks like it touches the axis may have a minimum of 0.02, and the reported minimum reveals it.
- Plot the difference, not just the functions. If you want to know where two expressions meet, plot each separately and look for the crossing, or plot their difference and look for its root. The second method converts an intersection question into a root question, which the summary reports numerically.
- Widen before concluding "no root". A root outside the window is invisible, and the summary only speaks about the interval you gave it.
- Use radians. All trigonometric functions take radians. To inspect a function defined in degrees, convert inside the expression, for example
sin(x * pi / 180). - Write powers with the caret.
x^2is a power;x2is read as an unknown name and produces an error message rather than a silent wrong answer. - Keep one expression per field. Long chained expressions still parse correctly, but splitting a comparison into three fields makes each curve individually legible in the legend and in the table.
This tool samples a function; it does not solve it. Roots and extrema are numerical estimates on a finite grid, not exact algebraic results, and a root where the curve merely touches the axis will not be reported at all. Anyone who needs an exact answer should use a symbolic method or a dedicated equation solver.
Only explicit functions of x are supported. A relation such as a circle written as x^2 + y^2 = 25 cannot be entered, because the tool needs a rule that produces one output per input. Implicit curves, parametric curves, polar equations and three-dimensional surfaces are outside its scope.
Vertical precision is limited by the automatic scaling. Extreme values near an asymptote are trimmed from the visible band so that the rest of the curve remains readable, which means the drawing never shows the full extent of a function that escapes to large magnitudes. The summary still reports the untrimmed minimum and maximum among the sampled points, and the two can differ substantially near a singularity.
Finally, the parser is a mathematical expression evaluator, not a programming language. There are no variables other than x, no assignment, no loops and no user-defined functions. That restriction is deliberate: it keeps the evaluation fast, predictable, and safe for input typed by strangers.
- ❓ What is a graphing calculator?
- ✅ It is a tool that draws the graph of a mathematical function by evaluating the expression at many input values and connecting the results. This one plots up to three explicit functions of x inside a window you choose and reports roots, extrema and undefined points for that window.
- ❓ How do I enter a power such as x squared?
- ✅ Use the caret, as in x^2. The parser evaluates powers right to left, so x^2^3 means x raised to the eighth power. Writing x2 without the caret produces an unknown-name error instead of a silent misreading.
- ❓ Can I write multiplication without a symbol?
- ✅ Yes. Expressions like 2x, 3(x + 1), x(x - 1) and 2sin(x) are all accepted, because the parser inserts the multiplication when a completed operand is followed by another operand or by an opening parenthesis.
- ❓ Why does the tool say a function has no roots when I can see a crossing?
- ✅ Either the crossing is outside the window, or the curve touches the axis without changing sign. The root detection compares the sign of neighbouring samples, so a tangency is invisible to it. Narrow the window around the point and compare the reported minimum with zero.
- ❓ Are trigonometric functions in degrees or radians?
- ✅ Radians. Every trigonometric function takes a radian argument, which is the convention in calculus and in programming. For degrees, multiply by pi and divide by 180 inside the expression, for example sin(x * pi / 180).
- ❓ Why does 1/x look like two separate curves?
- ✅ Because the function has no value at zero. When the sampling grid happens to land exactly on x = 0 the point is marked undefined and the curve visibly breaks; more often the grid steps just past zero, so the two branches are joined by a near-vertical segment that is a drawing artefact rather than a real value. Either way, treat the break as real and check x = 0 by hand. The same caution applies to tan(x) near its asymptotes.
- ❓ How accurate are the reported roots and extrema?
- ✅ They are numerical estimates on a grid of four hundred samples, with linear interpolation for crossings. Accuracy improves as you narrow the window, because the spacing between samples shrinks. They are never exact algebraic values.
- ❓ Can it plot a circle or an implicit equation?
- ✅ No. Only explicit functions of the form y equals an expression in x are supported. A circle would need two branches, an upper and a lower half, entered as separate expressions using the square root of a constant minus x squared.
- ❓ Why does sin(50x) look wrong?
- ✅ That function oscillates faster than the sampling grid can follow, a phenomenon called aliasing. With a window from -10 to 10 the samples are 0.05 apart and the period is about 0.126, so fewer than three samples land in each period. Narrow the window to fix it.
- ❓ Does the vertical scale change automatically?
- ✅ Yes. The visible vertical band is computed from the sampled values, with the top and bottom two percent trimmed so that a single asymptote does not flatten everything else. The summary still reports the untrimmed minimum and maximum.
- ❓ Can I save or share a graph?
- ✅ The plot can be inspected in place and the values table can be read or copied. The tool does not store graphs or create accounts, so each session starts fresh with the default expressions.
- ❓ Is this the same as a scientific calculator?
- ✅ No. A scientific calculator evaluates one expression and returns one number, while this tool evaluates an expression at hundreds of inputs and draws the relationship. For single-value computation use the Scientific Calculator, and for the rate of change of a curve use the Derivative Calculator.
References
- [1]Desmos. (n.d.). *Graphing Calculator*. https://www.desmos.com/calculator
- [2]Weisstein, E. W. (n.d.). *Function Graph*. Wolfram MathWorld. https://mathworld.wolfram.com/FunctionGraph.html
- [3]Abramson, J. (2021). *College Algebra 2e* (Chapter 3: Functions). OpenStax. https://openstax.org/books/college-algebra-2e/pages/3-1-functions-and-function-notation
- [4]NIST. (2024). *Digital Library of Mathematical Functions*. https://dlmf.nist.gov/
- [5]GeoGebra. (n.d.). *Graphing Calculator*. https://www.geogebra.org/graphing
- [6]Khan Academy. (n.d.). *Functions*. https://www.khanacademy.org/math/algebra/x2f8bb11595b61c86:functions
Last updated: November 13, 2026
UnByte — Independent Software Engineering
Every calculator references authoritative sources — Editorial policy

