Free math utilities for students, engineers, data analysts, and researchers. Compute matrix operations including transpose (swap rows and columns), determinant (a scalar describing the matrix's transformation), and inverse (the matrix that, multiplied by the original, yields the identity matrix). All math tools work with arbitrary precision and run entirely in your browser using JavaScript's native BigInt — there is no precision loss for large matrices or extreme values. Results include step-by-step working for educational use. Use cases include linear algebra homework, computer graphics transformations, machine learning preprocessing, and engineering calculations. Three tools cover the matrix operations that come up most often. Transposing a matrix is the first step in almost every linear algebra derivation and the first thing you reach for when reshaping a dataset for a model: it swaps rows and columns, so a 3×2 becomes a 2×3. If your data is arranged with one row per observation and one column per feature, transposing it gives you the feature-by-observation view that a matrix of weights is written in, and that mismatch is the single most common shape error in machine learning code. Transposing is also what turns a vector into a row or a column on demand, which is how you keep a dot product's operands in the orientation a particular formula expects. The transpose of a square matrix is the easy case; the interesting one is that transposing twice returns the original, and that the transpose of a product reverses the order of the factors. The determinant is the single scalar that tells you whether a transformation preserves area and whether a system of linear equations has a unique solution. For a 2×2 matrix it is ad minus bc; for anything larger the tool computes it by Gaussian elimination with partial pivoting, which is the same method you would use by hand. The value that matters more than the number itself is what it tells you about the matrix. A determinant of zero means the matrix is singular: it collapses a dimension, it has no inverse, and a system built on it either has no solution or infinitely many. For a 3×3 transformation, the determinant is the factor by which areas are scaled — a value of 2 means every area doubles, a value of 0.5 means every area halves, and a negative value means the transformation includes a reflection, which flips orientation as well as scaling. In engineering, a determinant of zero for a structural stiffness matrix is a mechanism rather than a stiff structure, and that is a design failure rather than a numerical curiosity. The inverse is the matrix that, multiplied by the original, yields the identity, so it undoes whatever the original did. The tool computes it with Gauss-Jordan elimination, and it detects a singular matrix and tells you so rather than returning nonsense. Knowing when an inverse exists is the useful part of the knowledge: any square matrix with a non-zero determinant has one, and a square matrix with a zero determinant does not. Where the inverse appears, it is usually because you are solving a system of equations — multiplying both sides of Ax = b by A⁻¹ gives x = A⁻¹b — or undoing a transformation you have already applied, or inverting a covariance matrix in statistics. The inverse is also the tool that reveals whether a problem is well-posed, which is why the singular case is reported explicitly instead of returning a large number full of error terms. Every step of the working is shown, not just the answer, so you can see which row or column a value came from and check it against your own notes. This is the difference between a tool that gives you a number and a tool that lets you learn the method, and it is the reason the elimination log exists rather than being decoration. When your answer disagrees with your textbook, the log is where the disagreement is visible: you can see the pivot row that was chosen, the row operation applied, and the point where the arithmetic diverges from the textbook's route. Nothing is uploaded: the matrix is parsed, computed and displayed entirely in the page, which means a matrix holding confidential figures never leaves your machine. Two honest limits are worth stating. These tools work on the numbers you type, so a matrix entered with the wrong dimensions will produce a wrong answer with no complaint — the transpose of a ragged array is not an error, it is just a different object, and nothing here can infer the shape you intended. And the arbitrary-precision arithmetic means exact answers for integers, which is what school and most coursework want, but it does not turn an approximation into an exact one; a determinant of 1/3 is not something a matrix of decimals can hold exactly, and the display shows you what the decimal expansion actually is. If you need additional math tools — logarithms, statistics, probability, an equation solver, or a unit converter, which already exists on LaiUse under Utility — use the Request a Tool page and we will consider adding it to the Math Tools category. One practical point about the elimination method itself, since it is what the logs actually show. Gaussian elimination reduces a matrix to upper-triangular form by eliminating entries below the diagonal, and partial pivoting means the algorithm swaps in the largest available value in the column before it eliminates. That detail is not cosmetic: without it, a matrix whose leading entry happens to be tiny can make the arithmetic amplify rounding error badly, and the result of a determinant can lose most of its significant digits. With pivoting, the growth factor stays bounded and the answer stays trustworthy. The Gauss-Jordan variant behind the inverse goes further, reducing the matrix all the way to the identity while carrying the right-hand side along, so the answer appears where the identity stands. For a 2×2 or 3×3 matrix you can check the answer by hand, and you should: the determinant of a 2×2 is ad minus bc, and the inverse of a 2×2 is one over the determinant times the matrix with its off-diagonal entries swapped, which takes seconds and catches an input typo immediately. For a 4×4 or larger, checking by hand stops being practical, and that is precisely the point of showing the working — the log is how you verify a result you cannot verify any other way. When an elimination step produces a row of zeros where a pivot should be, that is not a rounding artefact, it is the algorithm telling you the matrix is singular, and the tool reports it as such rather than returning an answer made of error terms.
Compute the determinant of any square matrix (N×N up to 10×10) via Gaussian elimination with partial pivoting. Includes step-by-step elimination log.
Find the inverse of any invertible square matrix using Gauss-Jordan elimination. Detects singular matrices and shows elimination steps.
Transpose any matrix — flip rows to columns. Enter values, see the result instantly, copy with one click. All processing local in your browser.