.NET API Getting Started

The Z3 .NET API provides a modern, object-oriented interface to Z3 for C#, F#, and other .NET languages.

Basic Workflow

A typical Z3 .NET program follows this pattern:

Create a Context (manages all Z3 objects)
Create expressions and constraints
Create a Solver and assert constraints
Check satisfiability
Extract results (models, proofs, etc.)
Dispose the context

Hello Z3

Let’s start with a minimal example:

using System;
using Microsoft.Z3;

class HelloZ3
{
    static void Main()
    {
        using (Context ctx = new Context())
        {
            Console.WriteLine($"Z3 version: {Microsoft.Z3.Version.ToString()}");
            Console.WriteLine("Context created successfully!");
        }
    }
}

The Context class implements IDisposable. Always use using statements or manually call Dispose() to prevent memory leaks.

Creating a Context

The Context is the central object that manages all Z3 operations:

using Microsoft.Z3;

// Basic context
using (Context ctx = new Context())
{
    // Use Z3...
}

// Context with custom configuration
var cfg = new Dictionary<string, string>
{
    { "model", "true" },       // Enable model generation
    { "proof", "false" },      // Disable proof generation
    { "timeout", "1000" }      // Set timeout to 1 second
};

using (Context ctx = new Context(cfg))
{
    // Use Z3...
}

Creating Expressions

Boolean Expressions

using (Context ctx = new Context())
{
    // Boolean constants
    BoolExpr bTrue = ctx.MkTrue();
    BoolExpr bFalse = ctx.MkFalse();
    
    // Boolean variables
    BoolExpr x = ctx.MkBoolConst("x");
    BoolExpr y = ctx.MkBoolConst("y");
    
    // Logical operations
    BoolExpr notX = ctx.MkNot(x);
    BoolExpr xAndY = ctx.MkAnd(x, y);
    BoolExpr xOrY = ctx.MkOr(x, y);
    BoolExpr xImpliesY = ctx.MkImplies(x, y);
    BoolExpr xIffY = ctx.MkIff(x, y);
}

Integer Expressions

using (Context ctx = new Context())
{
    // Integer constants
    IntExpr zero = ctx.MkInt(0);
    IntExpr five = ctx.MkInt(5);
    
    // Integer variables
    IntExpr x = ctx.MkIntConst("x");
    IntExpr y = ctx.MkIntConst("y");
    
    // Arithmetic operations
    ArithExpr xPlusY = ctx.MkAdd(x, y);
    ArithExpr xMinusY = ctx.MkSub(x, y);
    ArithExpr xTimesY = ctx.MkMul(x, y);
    IntExpr xDivY = ctx.MkDiv(x, y);
    IntExpr xModY = ctx.MkMod(x, y);
    
    // Comparisons
    BoolExpr xEqY = ctx.MkEq(x, y);
    BoolExpr xLtY = ctx.MkLt(x, y);
    BoolExpr xLeY = ctx.MkLe(x, y);
    BoolExpr xGtY = ctx.MkGt(x, y);
    BoolExpr xGeY = ctx.MkGe(x, y);
}

Real Expressions

using (Context ctx = new Context())
{
    // Real constants
    RealExpr half = ctx.MkReal(1, 2);      // 1/2
    RealExpr pi = ctx.MkReal("3.14159");
    
    // Real variables
    RealExpr x = ctx.MkRealConst("x");
    RealExpr y = ctx.MkRealConst("y");
    
    // Operations work similarly to integers
    ArithExpr sum = ctx.MkAdd(x, y);
    ArithExpr product = ctx.MkMul(x, y);
}

Using Solvers

Basic Solving

using System;
using Microsoft.Z3;

class BasicSolving
{
    static void Main()
    {
        using (Context ctx = new Context())
        {
            // Create variables
            IntExpr x = ctx.MkIntConst("x");
            IntExpr y = ctx.MkIntConst("y");
            
            // Create constraints: x > 0 AND y > x
            BoolExpr c1 = ctx.MkGt(x, ctx.MkInt(0));
            BoolExpr c2 = ctx.MkGt(y, x);
            
            // Create solver and add constraints
            Solver solver = ctx.MkSolver();
            solver.Assert(c1);
            solver.Assert(c2);
            
            // Check satisfiability
            Status result = solver.Check();
            
            if (result == Status.SATISFIABLE)
            {
                Console.WriteLine("SAT");
                Model model = solver.Model;
                Console.WriteLine($"Model: {model}");
                
                // Evaluate expressions in the model
                Console.WriteLine($"x = {model.Evaluate(x)}");
                Console.WriteLine($"y = {model.Evaluate(y)}");
            }
            else if (result == Status.UNSATISFIABLE)
            {
                Console.WriteLine("UNSAT");
            }
            else
            {
                Console.WriteLine("UNKNOWN");
            }
        }
    }
}

Incremental Solving with Push/Pop

using (Context ctx = new Context())
{
    IntExpr x = ctx.MkIntConst("x");
    Solver solver = ctx.MkSolver();
    
    // Add base constraint
    solver.Assert(ctx.MkGt(x, ctx.MkInt(0)));
    
    // Push: save state
    solver.Push();
    solver.Assert(ctx.MkLt(x, ctx.MkInt(5)));
    Console.WriteLine($"Check 1: {solver.Check()}");  // SAT: x in (0, 5)
    
    // Pop: restore state
    solver.Pop();
    
    // Add different constraint
    solver.Push();
    solver.Assert(ctx.MkGt(x, ctx.MkInt(100)));
    Console.WriteLine($"Check 2: {solver.Check()}");  // SAT: x > 100
    solver.Pop();
}

Complete Example: Proving De Morgan’s Law

Prove that ¬(x ∧ y) ⟺ (¬x ∨ ¬y):

using System;
using Microsoft.Z3;

class DeMorgan
{
    static void Main()
    {
        using (Context ctx = new Context())
        {
            // Create boolean variables
            BoolExpr x = ctx.MkBoolConst("x");
            BoolExpr y = ctx.MkBoolConst("y");
            
            // Build the formula: ¬(x ∧ y) ⟺ (¬x ∨ ¬y)
            BoolExpr leftSide = ctx.MkNot(ctx.MkAnd(x, y));
            BoolExpr rightSide = ctx.MkOr(ctx.MkNot(x), ctx.MkNot(y));
            BoolExpr conjecture = ctx.MkIff(leftSide, rightSide);
            
            // To prove, we show the negation is unsatisfiable
            BoolExpr negation = ctx.MkNot(conjecture);
            
            Solver solver = ctx.MkSolver();
            solver.Assert(negation);
            
            Status result = solver.Check();
            
            if (result == Status.UNSATISFIABLE)
            {
                Console.WriteLine("De Morgan's law is valid!");
            }
            else
            {
                Console.WriteLine("De Morgan's law is not valid (unexpected)");
            }
        }
    }
}

Complete Example: Solving Equations

Find values satisfying a system of equations:

using System;
using Microsoft.Z3;

class SolveEquations
{
    static void Main()
    {
        using (Context ctx = new Context())
        {
            // Variables
            IntExpr x = ctx.MkIntConst("x");
            IntExpr y = ctx.MkIntConst("y");
            
            // Equations:
            // x + y > 5
            // x - y < 2
            // x > 0
            BoolExpr eq1 = ctx.MkGt(ctx.MkAdd(x, y), ctx.MkInt(5));
            BoolExpr eq2 = ctx.MkLt(ctx.MkSub(x, y), ctx.MkInt(2));
            BoolExpr eq3 = ctx.MkGt(x, ctx.MkInt(0));
            
            Solver solver = ctx.MkSolver();
            solver.Assert(eq1, eq2, eq3);
            
            if (solver.Check() == Status.SATISFIABLE)
            {
                Model model = solver.Model;
                Console.WriteLine("Solution found:");
                Console.WriteLine($"x = {model.Evaluate(x)}");
                Console.WriteLine($"y = {model.Evaluate(y)}");
                
                // Verify
                var xVal = ((IntNum)model.Evaluate(x)).Int;
                var yVal = ((IntNum)model.Evaluate(y)).Int;
                Console.WriteLine("\nVerification:");
                Console.WriteLine($"x + y = {xVal + yVal} > 5? {xVal + yVal > 5}");
                Console.WriteLine($"x - y = {xVal - yVal} < 2? {xVal - yVal < 2}");
                Console.WriteLine($"x = {xVal} > 0? {xVal > 0}");
            }
            else
            {
                Console.WriteLine("No solution found");
            }
        }
    }
}

Working with Arrays

using (Context ctx = new Context())
{
    // Create array sort: Int -> Int
    Sort intSort = ctx.IntSort;
    ArraySort arraySort = ctx.MkArraySort(intSort, intSort);
    
    // Create array variable
    ArrayExpr a = (ArrayExpr)ctx.MkConst("a", arraySort);
    
    // Select element at index 0
    IntExpr elem0 = (IntExpr)ctx.MkSelect(a, ctx.MkInt(0));
    
    // Store value 42 at index 0
    ArrayExpr a2 = ctx.MkStore(a, ctx.MkInt(0), ctx.MkInt(42));
    
    // Create constraint: a2[0] == 42
    Solver solver = ctx.MkSolver();
    solver.Assert(ctx.MkEq(ctx.MkSelect(a2, ctx.MkInt(0)), ctx.MkInt(42)));
    
    Console.WriteLine(solver.Check());  // SATISFIABLE
}

Working with Quantifiers

using (Context ctx = new Context())
{
    // Create a universal quantifier: forall x. x >= 0 => x + 1 > 0
    IntExpr x = ctx.MkIntConst("x");
    BoolExpr body = ctx.MkImplies(
        ctx.MkGe(x, ctx.MkInt(0)),
        ctx.MkGt(ctx.MkAdd(x, ctx.MkInt(1)), ctx.MkInt(0))
    );
    
    BoolExpr forall = ctx.MkForall(
        new Expr[] { x },           // Quantified variables
        body,                        // Body
        1,                           // Weight
        null                         // Patterns
    );
    
    // Prove the formula
    Solver solver = ctx.MkSolver();
    solver.Assert(ctx.MkNot(forall));
    
    if (solver.Check() == Status.UNSATISFIABLE)
    {
        Console.WriteLine("Formula is valid!");
    }
}

Exception Handling

using System;
using Microsoft.Z3;

class ExceptionHandling
{
    static void Main()
    {
        try
        {
            using (Context ctx = new Context())
            {
                Solver solver = ctx.MkSolver();
                
                // Your Z3 operations
                IntExpr x = ctx.MkIntConst("x");
                solver.Assert(ctx.MkGt(x, ctx.MkInt(0)));
                
                Status result = solver.Check();
                Console.WriteLine($"Result: {result}");
            }
        }
        catch (Z3Exception ex)
        {
            Console.WriteLine($"Z3 error: {ex.Message}");
        }
        catch (Exception ex)
        {
            Console.WriteLine($"General error: {ex.Message}");
        }
    }
}

F# Example

Z3 works beautifully with F#:

open Microsoft.Z3
open System

[<EntryPoint>]
let main argv =
    use ctx = new Context()
    
    // Create variables
    let x = ctx.MkIntConst("x")
    let y = ctx.MkIntConst("y")
    
    // Create constraints: x + y > 5 AND x - y < 2
    let c1 = ctx.MkGt(ctx.MkAdd(x, y), ctx.MkInt(5))
    let c2 = ctx.MkLt(ctx.MkSub(x, y), ctx.MkInt(2))
    
    // Solve
    let solver = ctx.MkSolver()
    solver.Assert(c1, c2)
    
    match solver.Check() with
    | Status.SATISFIABLE ->
        printfn "SAT"
        printfn "Model: %O" solver.Model
        printfn "x = %O" (solver.Model.Evaluate(x))
        printfn "y = %O" (solver.Model.Evaluate(y))
    | Status.UNSATISFIABLE ->
        printfn "UNSAT"
    | _ ->
        printfn "UNKNOWN"
    
    0

Printing and Debugging

using (Context ctx = new Context())
{
    IntExpr x = ctx.MkIntConst("x");
    IntExpr y = ctx.MkIntConst("y");
    BoolExpr formula = ctx.MkGt(ctx.MkAdd(x, y), ctx.MkInt(5));
    
    // Print expressions
    Console.WriteLine($"Formula: {formula}");
    Console.WriteLine($"S-expression: {formula.SExpr()}");
    
    // Print solver state
    Solver solver = ctx.MkSolver();
    solver.Assert(formula);
    Console.WriteLine($"Solver: {solver}");
    
    // Print model
    if (solver.Check() == Status.SATISFIABLE)
    {
        Model model = solver.Model;
        Console.WriteLine($"Model: {model}");
        
        // Print all constants in the model
        foreach (FuncDecl decl in model.ConstDecls)
        {
            Console.WriteLine($"{decl.Name} = {model.ConstInterp(decl)}");
        }
    }
}

Using LINQ with Z3

You can use LINQ for cleaner code:

using System.Linq;
using Microsoft.Z3;

using (Context ctx = new Context())
{
    // Create multiple variables using LINQ
    var vars = Enumerable.Range(0, 5)
        .Select(i => ctx.MkIntConst($"x{i}"))
        .ToArray();
    
    // Create constraint: all different
    var distinct = ctx.MkDistinct(vars);
    
    // Create constraint: all in range [0, 4]
    var constraints = vars.Select(v => 
        ctx.MkAnd(
            ctx.MkGe(v, ctx.MkInt(0)),
            ctx.MkLe(v, ctx.MkInt(4))
        )
    ).ToArray();
    
    Solver solver = ctx.MkSolver();
    solver.Assert(distinct);
    solver.Assert(constraints);
    
    if (solver.Check() == Status.SATISFIABLE)
    {
        var model = solver.Model;
        foreach (var v in vars)
        {
            Console.WriteLine($"{v} = {model.Evaluate(v)}");
        }
    }
}

Best Practices

Always use using statements for Context to ensure proper disposal
Enable model generation if you need solutions: cfg["model"] = "true"
Use typed expressions (IntExpr, BoolExpr, etc.) for type safety
Check solver status before accessing Model property
Use Push/Pop for incremental solving to avoid recreating contexts
Handle Z3Exception in production code
Set timeouts for potentially long-running queries

Always check the solver status before accessing the Model property. Accessing it when the result is not SATISFIABLE will throw an exception.

Next Steps

API Reference

Explore the complete .NET API

.NET Examples

View more examples on GitHub

Z3 Guide

Interactive Z3 tutorial

NuGet Package

Official NuGet package

Python

C++

C

Java

.NET

JavaScript/TypeScript

Go

OCaml

Basic Workflow

Hello Z3

Creating a Context

Creating Expressions

Boolean Expressions

Integer Expressions

Real Expressions

Using Solvers

Basic Solving

Incremental Solving with Push/Pop

Complete Example: Proving De Morgan’s Law

Complete Example: Solving Equations

Working with Arrays

Working with Quantifiers

Exception Handling

F# Example

Printing and Debugging

Using LINQ with Z3

Best Practices

Next Steps

API Reference

.NET Examples

Z3 Guide

NuGet Package

Build docs developers (and LLMs) love

Python

C++

C

Java

.NET

JavaScript/TypeScript

Go

OCaml

​Basic Workflow

​Hello Z3

​Creating a Context

​Creating Expressions

​Boolean Expressions

​Integer Expressions

​Real Expressions

​Using Solvers

​Basic Solving

​Incremental Solving with Push/Pop

​Complete Example: Proving De Morgan’s Law

​Complete Example: Solving Equations

​Working with Arrays

​Working with Quantifiers

​Exception Handling

​F# Example

​Printing and Debugging

​Using LINQ with Z3

​Best Practices

​Next Steps

API Reference

.NET Examples

Z3 Guide

NuGet Package

Build docs developers (and LLMs) love

Basic Workflow

Hello Z3

Creating a Context

Creating Expressions

Boolean Expressions

Integer Expressions

Real Expressions

Using Solvers

Basic Solving

Incremental Solving with Push/Pop

Complete Example: Proving De Morgan’s Law

Complete Example: Solving Equations

Working with Arrays

Working with Quantifiers

Exception Handling

F# Example

Printing and Debugging

Using LINQ with Z3

Best Practices

Next Steps