Explain the difference between the low-level libdbus API and the high-level GDBus API?

The low-level libdbus API works directly with raw DBusMessage objects: you manually construct a message, append arguments one at a time with explicit type codes, send it, and then manually iterate the reply message's arguments back out, checking types as you go. It also requires manually integrating the library's file descriptors into whatever event loop your application uses, since libdbus doesn't assume any particular main loop.

DBusMessage *msg = dbus_message_new_method_call(
    "org.example.Service", "/org/example/Object",
    "org.example.Interface", "DoThing");
dbus_message_append_args(msg, DBUS_TYPE_STRING, &arg, DBUS_TYPE_INVALID);

GDBus's high-level API instead offers generated or dynamically created proxy objects that expose remote methods as ordinary function calls with native argument types, automatically maps signals and property changes onto GObject signals, and integrates natively with GLib's main loop and async patterns (GAsyncResult/callbacks or, in newer GLib, coroalike async functions).

The practical difference is development speed and safety versus control: low-level libdbus gives fine-grained control over exactly what's marshaled and when, useful for minimal-dependency code or unusual use cases, while GDBus's high-level layer eliminates most manual marshaling bugs and boilerplate for the overwhelming majority of ordinary client and service code.