Since C# 12, classes can have a primary constructor: parameters are declared next to the class name and are visible throughout the body. They are perfect for services with dependency injection.

Before and after

// Before
public class OrderService
{
    private readonly IOrderRepository _repo;
    private readonly ILogger<OrderService> _log;

    public OrderService(IOrderRepository repo, ILogger<OrderService> log)
    {
        _repo = repo;
        _log = log;
    }
}

// With the primary constructor
public class OrderService(IOrderRepository repo, ILogger<OrderService> log)
{
    public async Task<Order?> FindAsync(int id)
    {
        log.LogInformation("Looking up order {Id}", id);
        return await repo.GetAsync(id);
    }
}

The pitfall: parameters are not readonly

A primary constructor parameter can be modified inside the class. If you want the guarantee that it does not change, assign it to a readonly field:

public class Cache(TimeSpan duration)
{
    private readonly TimeSpan _duration = duration;
}

Other things to watch

• Do not use the same parameter both directly and to initialise a field: the compiler warns because two copies of the state would be created.
• For types that represent data, record types remain the best choice: they generate properties, equality and ToString.
• If the constructor must validate parameters, validation in a field initialiser works, but for complex logic a classic constructor is clearer.