diff --git a/libctru/include/3ds.h b/libctru/include/3ds.h
index a2c8e72..6d59617 100644
--- a/libctru/include/3ds.h
+++ b/libctru/include/3ds.h
@@ -16,6 +16,7 @@ extern "C" {
 #include <3ds/srv.h>
 #include <3ds/os.h>
 #include <3ds/synchronization.h>
+#include <3ds/thread.h>
 #include <3ds/gfx.h>
 #include <3ds/console.h>
 #include <3ds/env.h>
diff --git a/libctru/include/3ds/svc.h b/libctru/include/3ds/svc.h
index c5ab683..8a5af58 100644
--- a/libctru/include/3ds/svc.h
+++ b/libctru/include/3ds/svc.h
@@ -579,7 +579,7 @@ Result svcSetThreadIdealProcessor(Handle thread, s32 processorid);
  * @brief Returns the ID of the processor the current thread is running on.
  * @sa svcCreateThread
  */
-s32    svcGetProcessorID();
+s32    svcGetProcessorID(void);
 
 /**
  * @brief Gets the ID of a thread.
@@ -755,7 +755,7 @@ Result svcClearTimer(Handle timer);
  * @brief Gets the current system tick.
  * @return The current system tick.
  */
-u64    svcGetSystemTick();
+u64    svcGetSystemTick(void);
 ///@}
 
 ///@name System
diff --git a/libctru/include/3ds/thread.h b/libctru/include/3ds/thread.h
new file mode 100644
index 0000000..36ede63
--- /dev/null
+++ b/libctru/include/3ds/thread.h
@@ -0,0 +1,20 @@
+/**
+ * @file thread.h
+ * @brief Provides functions to use threads.
+ */
+#pragma once
+#include <3ds/types.h>
+#include <3ds/result.h>
+#include <3ds/synchronization.h>
+#include <3ds/svc.h>
+
+typedef struct Thread_tag* Thread;
+
+Thread threadCreate(ThreadFunc entrypoint, void* arg, size_t stack_size, int prio, int affinity, bool detached);
+Handle threadGetHandle(Thread thread);
+int threadGetExitCode(Thread thread);
+void threadFree(Thread thread);
+Result threadJoin(Thread thread, u64 timeout_ns);
+
+Thread threadGetCurrent(void);
+void threadExit(int rc) __attribute__((noreturn));
diff --git a/libctru/source/internal.h b/libctru/source/internal.h
index 53a4c9f..269c9c3 100644
--- a/libctru/source/internal.h
+++ b/libctru/source/internal.h
@@ -1,8 +1,9 @@
 #pragma once
+#include <sys/reent.h>
 #include <3ds/types.h>
 #include <3ds/result.h>
 #include <3ds/svc.h>
-#include <sys/reent.h>
+#include <3ds/thread.h>
 
 #define THREADVARS_MAGIC  0x21545624 // !TV$
 #define FS_OVERRIDE_MAGIC 0x21465324 // !FS$
@@ -13,6 +14,9 @@ typedef struct
 	// Magic value used to check if the struct is initialized
 	u32 magic;
 
+	// Pointer to the current thread (if exists)
+	Thread thread_ptr;
+
 	// Pointer to this thread's newlib state
 	struct _reent* reent;
 
diff --git a/libctru/source/system/syscalls.c b/libctru/source/system/syscalls.c
index 3cb5734..a27482b 100644
--- a/libctru/source/system/syscalls.c
+++ b/libctru/source/system/syscalls.c
@@ -45,4 +45,5 @@ void __system_initSyscalls(void)
 	ThreadVars* tv = getThreadVars();
 	tv->magic = THREADVARS_MAGIC;
 	tv->reent = _impure_ptr;
+	tv->thread_ptr = NULL;
 }
diff --git a/libctru/source/thread.c b/libctru/source/thread.c
new file mode 100644
index 0000000..382534c
--- /dev/null
+++ b/libctru/source/thread.c
@@ -0,0 +1,112 @@
+#include "internal.h"
+#include <stdlib.h>
+#include <string.h>
+
+struct Thread_tag
+{
+	Handle handle;
+	ThreadFunc ep;
+	void* arg;
+	int rc;
+	bool detached, finished;
+	struct _reent reent;
+	void* stacktop;
+};
+
+static void __panic(void)
+{
+	svcBreak(USERBREAK_PANIC);
+	for (;;);
+}
+
+static void _thread_begin(void* arg)
+{
+	Thread t = (Thread)arg;
+	ThreadVars* tv = getThreadVars();
+	tv->magic = THREADVARS_MAGIC;
+	tv->reent = &t->reent;
+	tv->thread_ptr = t;
+	t->ep(t->arg);
+	threadExit(0);
+}
+
+Thread threadCreate(ThreadFunc entrypoint, void* arg, size_t stack_size, int prio, int affinity, bool detached)
+{
+	size_t stackoffset = (sizeof(struct Thread_tag)+7)&~7;
+	size_t allocsize   = stackoffset + ((stack_size+7)&~7);
+	if (allocsize < stackoffset) return NULL; // guard against overflow
+	if ((allocsize-stackoffset) < stack_size) return NULL; // guard against overflow
+	Thread t = (Thread)malloc(allocsize);
+	if (!t) return NULL;
+
+	t->ep       = entrypoint;
+	t->arg      = arg;
+	t->detached = detached;
+	t->finished = false;
+	t->stacktop = (u8*)t + allocsize;
+
+	// Set up child thread's reent struct, inheriting standard file handles
+	_REENT_INIT_PTR(&t->reent);
+	struct _reent* cur = getThreadVars()->reent;
+	t->reent._stdin  = cur->_stdin;
+	t->reent._stdout = cur->_stdout;
+	t->reent._stderr = cur->_stderr;
+
+	Result rc;
+	rc = svcCreateThread(&t->handle, _thread_begin, (u32)t, (u32*)t->stacktop, prio, affinity);
+	if (R_FAILED(rc))
+	{
+		free(t);
+		return NULL;
+	}
+
+	return t;
+}
+
+Handle threadGetHandle(Thread thread)
+{
+	if (!thread || thread->finished) return ~0UL;
+	return thread->handle;
+}
+
+int threadGetExitCode(Thread thread)
+{
+	if (!thread || !thread->finished) return 0;
+	return thread->rc;
+}
+
+void threadFree(Thread thread)
+{
+	if (!thread || !thread->finished) return;
+	svcCloseHandle(thread->handle);
+	free(thread);
+}
+
+Result threadJoin(Thread thread, u64 timeout_ns)
+{
+	if (!thread || thread->finished) return 0;
+	return svcWaitSynchronization(thread->handle, timeout_ns);
+}
+
+Thread threadGetCurrent(void)
+{
+	ThreadVars* tv = getThreadVars();
+	if (tv->magic != THREADVARS_MAGIC)
+		__panic();
+	return tv->thread_ptr;
+}
+
+void threadExit(int rc)
+{
+	Thread t = threadGetCurrent();
+	if (!t)
+		__panic();
+
+	t->finished = true;
+	if (t->detached)
+		threadFree(t);
+	else
+		t->rc = rc;
+
+	svcExitThread();
+}