Maintained demo · Constraint solving

Sudoku Constraint Solver

Play a generated 9×9 puzzle, validate your entries, or open Developer View to inspect propagation, minimum-remaining-values branching, and backtracking.

Sudoku board

How it works

Constraint propagation + search

The solver reduces each cell’s domain before branching. Search is used only when propagation cannot determine the next value.

Domains

Every empty cell tracks digits not already used in its row, column, or 3×3 box.

Propagation

Naked and hidden singles reduce domains until no more forced assignments remain.

MRV backtracking

The solver branches on the cell with the fewest candidates, tries a value, and reverses the choice if it reaches a contradiction.

Complexity and limits

Generalized Sudoku solving is NP-complete. Backtracking has exponential worst-case behavior, although propagation and MRV reduce practical search. A solved board demonstrates algorithm behavior, not puzzle difficulty calibration.