diff --git a/CMakeLists.txt b/CMakeLists.txt
index 92b45615a..4db24fbbb 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -239,7 +239,9 @@ target_link_libraries(llama PRIVATE utils ggml ${LLAMA_EXTRA_LIBS})
 # Executables
 #
 
-add_executable(main run.cpp)
+add_executable(main
+               main.cpp
+               run.cpp)
 target_link_libraries(main PRIVATE llama ggml utils)
 
 add_executable(quantize quantize.cpp)
diff --git a/Makefile b/Makefile
index a8f09db7a..2f11ea166 100644
--- a/Makefile
+++ b/Makefile
@@ -226,11 +226,14 @@ llama.o: llama.cpp llama.h
 utils.o: utils.cpp utils.h
 	$(CXX) $(CXXFLAGS) -c utils.cpp -o utils.o
 
+run.o: run.cpp run.h
+	$(CXX) $(CXXFLAGS) -c run.cpp -o run.o
+
 clean:
 	rm -f *.o main quantize
 
-main: run.cpp ggml.o llama.o utils.o
-	$(CXX) $(CXXFLAGS) run.cpp ggml.o llama.o utils.o -o main $(LDFLAGS)
+main: main.cpp ggml.o llama.o utils.o run.o
+	$(CXX) $(CXXFLAGS) main.cpp ggml.o llama.o utils.o run.o -o main $(LDFLAGS)
 	@echo "\x1b[36mrun ./main -h for help\x1b[0m"
 
 quantize: quantize.cpp ggml.o llama.o utils.o
diff --git a/main.cpp b/main.cpp
new file mode 100644
index 000000000..61fec449a
--- /dev/null
+++ b/main.cpp
@@ -0,0 +1,5 @@
+#include "run.h"
+
+int main(int argc, char ** argv) {
+    return run(argc, argv);
+}
diff --git a/run.cpp b/run.cpp
index 4569ef2a1..e0db76974 100644
--- a/run.cpp
+++ b/run.cpp
@@ -154,7 +154,7 @@ void sigint_handler(int signo) {
 }
 #endif
 
-int main(int argc, char ** argv) {
+int run(int argc, char ** argv) {
     // has to be called once at the start of the program to init ggml stuff
     ggml_time_init();
 
diff --git a/run.h b/run.h
new file mode 100644
index 000000000..4a490bb98
--- /dev/null
+++ b/run.h
@@ -0,0 +1,3 @@
+#pragma once
+
+int run(int argc, char ** argv);