diff --git a/src/lib/nbt.cpp b/src/lib/nbt.cpp
index 3935af6..4642077 100644
--- a/src/lib/nbt.cpp
+++ b/src/lib/nbt.cpp
@@ -42,7 +42,7 @@ int endianness_example() {
 }
 
 namespace NBT {
-    namespace helpers {
+    namespace helper {
         ErrorOr<int8_t> readInt8(uint8_t* data[], uint64_t dataSize, uint64_t currentPosition) {
             //TODO: implement
             return ErrorOr<int8_t>(0);
diff --git a/src/lib/nbt.h++ b/src/lib/nbt.h++
index 2c9d175..1ca5cfb 100644
--- a/src/lib/nbt.h++
+++ b/src/lib/nbt.h++
@@ -40,5 +40,31 @@
 #include "error.h++"
 
 namespace NBT {
+    namespace helper {
+        ErrorOr<int8_t> readInt8(uint8_t* data[], uint64_t dataSize, uint64_t currentPosition);
+
+        ErrorOr<int16_t> readInt16(uint8_t* data[], uint64_t dataSize, uint64_t currentPosition);
+
+        ErrorOr<int32_t> readInt32(uint8_t* data[], uint64_t dataSize, uint64_t currentPosition);
+
+        ErrorOr<int64_t> readInt64(uint8_t* data[], uint64_t dataSize, uint64_t currentPosition);
+
+        //FIXME: we just assume that float is a single-precision IEEE754
+        // floating point number
+        ErrorOr<float> readFloat32(uint8_t* data[], uint64_t dataSize, uint64_t currentPosition);
+
+        //FIXME: we just assume that double is a double-precision IEEE754
+        // floating point number
+        ErrorOr<double> readFloat64(uint8_t* data[], uint64_t dataSize, uint64_t currentPosition);
+
+        ErrorOr<std::vector<int8_t>> readInt8Array(uint8_t* data[], uint64_t dataSize, uint64_t currentPosition);
+
+        //ErrorOr<> readString(uint8_t* data[], uint64_t dataSize, uint64_t currentPosition);
+
+        ErrorOr<std::vector<int32_t>> readInt32Array(uint8_t* data[], uint64_t dataSize, uint64_t currentPosition);
+
+        ErrorOr<std::vector<int64_t>> readInt64Array(uint8_t* data[], uint64_t dataSize, uint64_t currentPosition);
+    }
+
     bool validateRawNBTData(uint8_t* data[], int length);
 }