diff --git a/src/tz/core/messaging.hpp b/src/tz/core/messaging.hpp
index 8a174bfb3e..4fc5c73563 100644
--- a/src/tz/core/messaging.hpp
+++ b/src/tz/core/messaging.hpp
@@ -4,6 +4,7 @@
 #include <vector>
 #include <mutex>
 #include <ranges>
+#include <variant>
 
 namespace tz
 {
@@ -23,9 +24,9 @@ namespace tz
 		struct with_lock
 		{
 		protected:
-			std::mutex mtx = {};	
+			mutable std::mutex mtx = {};	
 			using lock_t = std::unique_lock<std::mutex>;
-			lock_t lock() {return lock_t{this->mtx};}
+			lock_t lock() const {return lock_t{this->mtx};}
 		};
 	}
 
@@ -47,13 +48,13 @@ namespace tz
 		std::size_t message_count() const;
 	protected:
 		// Processes all messages and then clears the list. Thread-safe.
-		void process_messages();
+		virtual void process_messages();
 		/// Invoked when a message is sent. Your override can return false to drop the message. By default, no messages are dropped. Not thread safe.
 		virtual bool on_send_message(const M& msg) {return true;};
 		/// Invoked when a message is processed. You must override this. Not thread safe.
 		/// @note If processing a message causes the receiver to send a new message to itself, this is safe, but will not be processed until the next invocation to `process_messages`.
 		virtual void process_message(const M& msg) = 0;
-	private:
+	protected:
 		using base_t = std::conditional_t<thread_safe, detail::with_lock, detail::no_lock>;
 		std::vector<M> messages = {};
 	};
@@ -73,16 +74,23 @@ namespace tz
 		/// Create a passer with no target.
 		message_passer() = default;
 		/// Create a passer with an existing target.
-		message_passer(message_receiver<M, thread_safe>& target);
+		message_passer(message_receiver<M, false>& target);
+		/// Create a passer with an existing target.
+		message_passer(message_receiver<M, true>& target);
+		/// Redirect all messages to the target.
+		virtual void process_messages() override final;
 		/// Redirect messages to the target.
 		virtual void process_message(const M& msg) override final;
 
 		/// Redirect messages to a new target. Not thread-safe.
-		void set_target(message_receiver<M, thread_safe>& target);
+		void set_target(message_receiver<M, true>& target);
+		/// Redirect messages to a new target. Not thread-safe.
+		void set_target(message_receiver<M, false>& target);
 		/// Clear the target. Processed messages will no longer be redirected. Not thread-safe.
 		void clear_target();
 	private:
-		message_receiver<M, thread_safe>* target = nullptr;
+		using base_t = message_receiver<M, thread_safe>::base_t;
+		std::variant<message_receiver<M, true>*, message_receiver<M, false>*, std::monostate> target = std::monostate{};
 	};
 }
 
diff --git a/src/tz/core/messaging.inl b/src/tz/core/messaging.inl
index 38cd982053..0fc3c842ef 100644
--- a/src/tz/core/messaging.inl
+++ b/src/tz/core/messaging.inl
@@ -60,21 +60,59 @@ namespace tz
 	}
 
 	template<tz::message M, bool thread_safe>
-	message_passer<M, thread_safe>::message_passer(message_receiver<M, thread_safe>& parent):
-	parent(&parent)
+	message_passer<M, thread_safe>::message_passer(message_receiver<M, true>& target):
+	target(&target)
 	{}
 
 	template<tz::message M, bool thread_safe>
-	void message_passer<M, thread_safe>::process_message(const M& msg)
+	message_passer<M, thread_safe>::message_passer(message_receiver<M, false>& target):
+	target(&target)
+	{}
+
+	template<tz::message M, bool thread_safe>
+	void message_passer<M, thread_safe>::process_messages()
 	{
-		if(this->target != nullptr)
+		typename base_t::lock_t lock = base_t::lock();
+		if(this->messages.empty())
 		{
-			this->target->send_message(msg);
+			return;
 		}
+		std::visit([this](auto&& receiver_ptr)
+		{
+			if constexpr(!std::is_same_v<std::decay_t<decltype(receiver_ptr)>, std::monostate>)
+			{
+				if(receiver_ptr != nullptr)
+				{
+					receiver_ptr->send_messages(this->messages);
+				}
+			}
+		}, this->target);
+		this->messages.clear();
+	}
+
+	template<tz::message M, bool thread_safe>
+	void message_passer<M, thread_safe>::process_message(const M& msg)
+	{
+		std::visit([&msg](auto&& receiver_ptr)
+		{
+			if constexpr(!std::is_same_v<std::decay_t<decltype(receiver_ptr)>, std::monostate>)
+			{
+				if(receiver_ptr != nullptr)
+				{
+					receiver_ptr->send_message(msg);
+				}
+			}
+		}, this->target);
+	}
+
+	template<tz::message M, bool thread_safe>
+	void message_passer<M, thread_safe>::set_target(message_receiver<M, true>& target)
+	{
+		this->target = &target;
 	}
 
 	template<tz::message M, bool thread_safe>
-	void message_passer<M, thread_safe>::set_target(message_receiver<M, thread_safe>& target)
+	void message_passer<M, thread_safe>::set_target(message_receiver<M, false>& target)
 	{
 		this->target = &target;
 	}
@@ -82,6 +120,6 @@ namespace tz
 	template<tz::message M, bool thread_safe>
 	void message_passer<M, thread_safe>::clear_target()
 	{
-		this->target = nullptr;
+		this->target = std::monostate{};
 	}
 }
diff --git a/test/core/message_test.cpp b/test/core/message_test.cpp
index 872da174a9..681bd299fa 100644
--- a/test/core/message_test.cpp
+++ b/test/core/message_test.cpp
@@ -8,7 +8,7 @@ struct assignment_msg
 };
 static_assert(tz::message<assignment_msg>);
 
-class deferred_assignment : public tz::message_receiver<assignment_msg>
+class deferred_assignment : public tz::message_receiver<assignment_msg, true>
 {
 public:
 	int x = 0;
@@ -93,8 +93,46 @@ void scenario2()
 	tz::assert(sys.x == 95);
 }
 
+// scenario 3 - thread-safe and non-thread-safe message passers to the same thread-safe message receiver.
+
+template<bool thread_safe>
+struct deferred_assignment_passer : public tz::message_passer<assignment_msg, thread_safe>
+{
+	using tz::message_passer<assignment_msg, thread_safe>::message_passer;
+	void update()
+	{
+		tz::message_passer<assignment_msg, thread_safe>::process_messages();
+	}
+};
+
+using deferred_assignment_passer_ts = deferred_assignment_passer<true>;
+using deferred_assignment_passer_nts = deferred_assignment_passer<false>;
+
+void scenario3()
+{
+	deferred_assignment sys;
+
+	deferred_assignment_passer_ts ts(sys);
+	ts.set_target(sys);
+
+	deferred_assignment_passer_nts nts(sys);
+	nts.set_target(sys);
+
+	tz::assert(sys.x == 0);
+	ts.send_message({.new_val = 2});
+	nts.send_message({.new_val = 5});
+	// both of the following updates can happen concurrently on different threads, and its safe (so long as `nts` isnt also being sent new messages at the same time)
+	ts.update();
+	nts.update();
+	tz::assert(sys.message_count() == 2);
+	tz::assert(sys.x == 0);
+	sys.update();
+	tz::assert(sys.x == 5);
+}
+
 int main()
 {
 	scenario1();
 	scenario2();
+	scenario3();
 }