diff --git a/templates/gapuino/include/api.h b/templates/gapuino/include/api.h
new file mode 100644
index 0000000..0b1d09b
--- /dev/null
+++ b/templates/gapuino/include/api.h
@@ -0,0 +1,5 @@
+#define CRYPTO_KEYBYTES 16
+#define CRYPTO_NSECBYTES 0
+#define CRYPTO_NPUBBYTES 12
+#define CRYPTO_ABYTES 8
+#define CRYPTO_NOOVERLAP 1
diff --git a/templates/gapuino/include/uartp.h b/templates/gapuino/include/uartp.h
index 3a80a60..085073e 100644
--- a/templates/gapuino/include/uartp.h
+++ b/templates/gapuino/include/uartp.h
@@ -1,3 +1,5 @@
+#include <stdint.h>
+
 #pragma once
 #ifdef __cplusplus
 extern "C" {
diff --git a/templates/gapuino/platformio.ini b/templates/gapuino/platformio.ini
index 54a9abc..e614720 100644
--- a/templates/gapuino/platformio.ini
+++ b/templates/gapuino/platformio.ini
@@ -12,4 +12,6 @@
 platform = riscv_gap
 board = gapuino
 framework = mbed
+board_upload.boot_mode = jtag_hyper
+board_upload.commands = reqloop ioloop start wait
 
diff --git a/templates/gapuino/src/api.h b/templates/gapuino/src/api.h
new file mode 100644
index 0000000..290a70d
--- /dev/null
+++ b/templates/gapuino/src/api.h
@@ -0,0 +1,5 @@
+#define CRYPTO_KEYBYTES  32
+#define CRYPTO_NSECBYTES 0
+#define CRYPTO_NPUBBYTES 12
+#define CRYPTO_ABYTES 8
+#define CRYPTO_NOOVERLAP 1
diff --git a/templates/gapuino/src/encrypt.c b/templates/gapuino/src/encrypt.c
new file mode 100644
index 0000000..2e661d5
--- /dev/null
+++ b/templates/gapuino/src/encrypt.c
@@ -0,0 +1,199 @@
+/*
+     TinyJAMBU-256: 256-bit key, 96-bit IV 
+	 Reference Implementation for 32-bit processor 
+     The state consists of four 32-bit registers       
+     state[3] || state[2] || state[1] || state[0]
+
+	 Implemented by Hongjun Wu
+*/   
+
+#include <string.h>  
+#include <stdio.h> 
+#include "crypto_aead.h"
+
+#define FrameBitsIV  0x10  
+#define FrameBitsAD  0x30  
+#define FrameBitsPC  0x50  //Framebits for plaintext/ciphertext      
+#define FrameBitsFinalization 0x70       
+
+#define NROUND1 128*3 
+#define NROUND2 128*10
+
+/*non-optimized state update function*/    
+void state_update(unsigned int *state, const unsigned char *key, unsigned int number_of_steps)
+{
+	unsigned int i; 
+	unsigned int t1, t2, t3, t4, feedback;
+	//in each iteration, we compute 256 steps of the state update function. 
+	for (i = 0; i < (number_of_steps >> 5); i++)
+	{
+		t1 = (state[1] >> 15) | (state[2] << 17);  // 47 = 1*32+15 
+		t2 = (state[2] >> 6)  | (state[3] << 26);  // 47 + 23 = 70 = 2*32 + 6 
+		t3 = (state[2] >> 21) | (state[3] << 11);  // 47 + 23 + 15 = 85 = 2*32 + 21      
+		t4 = (state[2] >> 27) | (state[3] << 5);   // 47 + 23 + 15 + 6 = 91 = 2*32 + 27 
+		feedback = state[0] ^ t1 ^ (~(t2 & t3)) ^ t4 ^ ((unsigned int*)key)[i & 7];
+		// shift 32 bit positions 
+		state[0] = state[1]; state[1] = state[2]; state[2] = state[3];
+		state[3] = feedback;
+	}
+}
+  
+// The initialization  
+/* The input to initialization is the 128-bit key; 96-bit IV;*/
+void initialization(const unsigned char *key, const unsigned char *iv, unsigned int *state)
+{
+        int i;
+
+        //initialize the state as 0  
+		for (i = 0; i < 4; i++) state[i] = 0;     
+
+		//update the state with the key  
+		state_update(state, key, NROUND2);  
+
+		//introduce IV into the state  
+        for (i = 0;  i < 3; i++)  
+        {
+			state[1] ^= FrameBitsIV;   
+			state_update(state, key, NROUND1); 
+			state[3] ^= ((unsigned int*)iv)[i]; 
+		}   
+}
+
+//process the associated data   
+void process_ad(const unsigned char *k, const unsigned char *ad, unsigned long long adlen, unsigned int *state)
+{
+	unsigned long long i; 
+	unsigned int j; 
+
+	for (i = 0; i < (adlen >> 2); i++)
+	{
+		state[1] ^= FrameBitsAD;
+		state_update(state, k, NROUND1);
+		state[3] ^= ((unsigned int*)ad)[i];
+	}
+
+	// if adlen is not a multiple of 4, we process the remaining bytes
+	if ((adlen & 3) > 0)
+	{
+		state[1] ^= FrameBitsAD;
+		state_update(state, k, NROUND1);
+		for (j = 0; j < (adlen & 3); j++)  ((unsigned char*)state)[12 + j] ^= ad[(i << 2) + j];
+		state[1] ^= adlen & 3;
+	}   
+}     
+
+//encrypt a message  
+int crypto_aead_encrypt(
+	unsigned char *c,unsigned long long *clen,
+	const unsigned char *m,unsigned long long mlen,
+	const unsigned char *ad,unsigned long long adlen,
+	const unsigned char *nsec,
+	const unsigned char *npub,
+	const unsigned char *k
+	)
+{
+    unsigned long long i;
+	unsigned int j; 
+    unsigned char mac[8]; 
+    unsigned int state[4];   
+
+    //initialization stage
+    initialization(k, npub, state);
+
+    //process the associated data   
+	process_ad(k, ad, adlen, state); 
+
+	//process the plaintext    
+	for (i = 0; i < (mlen >> 2); i++)
+	{
+		state[1] ^= FrameBitsPC;  
+		state_update(state, k, NROUND2); 
+		state[3] ^= ((unsigned int*)m)[i];  
+		((unsigned int*)c)[i] = state[2] ^ ((unsigned int*)m)[i];  
+	}
+	// if mlen is not a multiple of 4, we process the remaining bytes
+	if ((mlen & 3) > 0)
+	{   
+		state[1] ^= FrameBitsPC; 
+		state_update(state, k, NROUND2);    
+		for (j = 0; j < (mlen & 3); j++)  
+		{
+			((unsigned char*)state)[12 + j] ^= m[(i << 2) + j];   
+			c[(i << 2) + j] = ((unsigned char*)state)[8 + j] ^ m[(i << 2) + j];
+		}   
+		state[1] ^= mlen & 3;   
+	}
+
+	//finalization stage, we assume that the tag length is 8 bytes
+	state[1] ^= FrameBitsFinalization;
+	state_update(state, k, NROUND2);
+	((unsigned int*)mac)[0] = state[2];
+
+	state[1] ^= FrameBitsFinalization;
+	state_update(state, k, NROUND1);
+	((unsigned int*)mac)[1] = state[2];
+
+    *clen = mlen + 8; 
+    memcpy(c + mlen, mac, 8);  
+
+    return 0;
+}
+
+//decrypt a message
+int crypto_aead_decrypt(
+	unsigned char *m,unsigned long long *mlen,
+	unsigned char *nsec,
+	const unsigned char *c,unsigned long long clen,
+	const unsigned char *ad,unsigned long long adlen,
+	const unsigned char *npub,
+	const unsigned char *k
+	)
+{
+	unsigned long long i;
+	unsigned int j, check = 0;
+ 	unsigned char mac[8];
+	unsigned int state[4];
+
+	*mlen = clen - 8; 
+
+	//initialization stage
+	initialization(k, npub, state);
+
+	//process the associated data   
+	process_ad(k, ad, adlen, state);
+
+	//process the ciphertext    
+	for (i = 0; i < (*mlen >> 2); i++)
+	{
+		state[1] ^= FrameBitsPC;
+		state_update(state, k, NROUND2);
+		((unsigned int*)m)[i] = state[2] ^ ((unsigned int*)c)[i];
+		state[3] ^= ((unsigned int*)m)[i]; 
+ 	}
+	// if mlen is not a multiple of 4, we process the remaining bytes
+	if ((*mlen & 3) > 0)   
+	{
+		state[1] ^= FrameBitsPC;  
+		state_update(state, k, NROUND2);
+		for (j = 0; j < (*mlen & 3); j++)
+		{
+			m[(i << 2) + j] = c[(i << 2) + j] ^ ((unsigned char*)state)[8 + j];
+			((unsigned char*)state)[12 + j] ^= m[(i << 2) + j];
+		}   
+		state[1] ^= *mlen & 3;  
+	}
+	
+	//finalization stage, we assume that the tag length is 8 bytes
+	state[1] ^= FrameBitsFinalization;
+	state_update(state, k, NROUND2);
+	((unsigned int*)mac)[0] = state[2];
+
+	state[1] ^= FrameBitsFinalization;
+	state_update(state, k, NROUND1);
+	((unsigned int*)mac)[1] = state[2];
+
+	//verification of the authentication tag   
+	for (j = 0; j < 8; j++) { check |= (mac[j] ^ c[clen - 8 + j]); }
+    if (check == 0) return 0;  
+    else return -1;
+}
diff --git a/templates/gapuino/src/main.ino b/templates/gapuino/src/main.cpp
similarity index 96%
rename from templates/gapuino/src/main.ino
rename to templates/gapuino/src/main.cpp
index c420b11..86928e1 100644
--- a/templates/gapuino/src/main.ino
+++ b/templates/gapuino/src/main.cpp
@@ -1,6 +1,6 @@
-#include "crypto_aead.h"
 #include "api.h"
 #include "uartp.h"
+#include "crypto_aead.h"
 #include "mbed.h"
 
 #define MAX_BYTES 100
@@ -23,7 +23,8 @@ unsigned long long int clen = 0;
 int res = 0;
 
 void setup();
-void loop();
+//void loop();
+DigitalOut crypto_pin(CRYPTO_BUSY);
 
 #ifdef __cplusplus
 extern "C" {
@@ -49,12 +50,11 @@ void my_assert(bool b) {
     if (b)
         return;
     for(;;)
-        yield();
+        wait(1);
 }
 
 void setup() {
     device.baud(115200);
-    DigitalOut crypto_pin(CRYPTO_BUSY);
     crypto_pin = 1;
     wait(100);
     memset(npub, 0, CRYPTO_NPUBBYTES);
@@ -66,12 +66,12 @@ void setup() {
     device.printf("Hello, World!");
 }
 
-void loop() {
+int main() {
     int res;
     uint16_t len = uartp_recv(cmdbuf, CMDBUF_LEN - 1);
     uint8_t action = cmdbuf[0];
     if (len == 0 || len > CMDBUF_LEN - 1)
-        return;
+        return 1;
 
     uint16_t l = len - 1;
     uint16_t rl = 0;
diff --git a/templates/gapuino/test b/templates/gapuino/test
index 35a54c3..09f0b7d 100755
--- a/templates/gapuino/test
+++ b/templates/gapuino/test
@@ -28,7 +28,7 @@ def get_serial():
     devices = [ p.device for p in ports ]
     devices.sort()
     # Might also be 0, one is JTAG, one UART
-    return devices[-1]
+    return devices[0]
 
 
 class UARTP: