diff --git a/include/duckhandy/endianness.hpp b/include/duckhandy/endianness.hpp
new file mode 100644
index 0000000..0da8d0c
--- /dev/null
+++ b/include/duckhandy/endianness.hpp
@@ -0,0 +1,179 @@
+/* Copyright 2016, 2017 Michele Santullo
+ * This file is part of "duckhandy".
+ *
+ * "duckhandy" is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * "duckhandy" is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with "duckhandy".  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef id1A975372553B45BC8C4E42CDBDD97497
+#define id1A975372553B45BC8C4E42CDBDD97497
+
+#include <cstdint>
+#include <endian.h>
+
+namespace dhandy {
+	template <typename T> constexpr T htobe (T parV);
+	template <typename T> constexpr T htole (T parV);
+	template <typename T> constexpr T betoh (T parV);
+	template <typename T> constexpr T letoh (T parV);
+
+	template <>
+	[[gnu::pure]] inline constexpr uint8_t htobe (uint8_t parV) {
+		return parV;
+	}
+	template <>
+	[[gnu::pure]] inline constexpr uint16_t htobe (uint16_t parV) {
+#if __BYTE_ORDER == __BIG_ENDIAN
+		return parV;
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+		return __builtin_bswap16(parV);
+#else
+#	error "Unsupported endianness"
+#endif
+	}
+	template <>
+	[[gnu::pure]] inline constexpr uint32_t htobe (uint32_t parV) {
+#if __BYTE_ORDER == __BIG_ENDIAN
+		return parV;
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+		return __builtin_bswap32(parV);
+#else
+#	error "Unsupported endianness"
+#endif
+	}
+	template <>
+	[[gnu::pure]] inline constexpr uint64_t htobe (uint64_t parV) {
+#if __BYTE_ORDER == __BIG_ENDIAN
+		return parV;
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+		return __builtin_bswap64(parV);
+#else
+#	error "Unsupported endianness"
+#endif
+	}
+	template <>
+	[[gnu::pure]] inline constexpr int8_t htobe (int8_t parV) {
+		return parV;
+	}
+	template <>
+	[[gnu::pure]] inline constexpr int16_t htobe (int16_t parV) {
+#if __BYTE_ORDER == __BIG_ENDIAN
+		return parV;
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+		return static_cast<int16_t>(__builtin_bswap16(static_cast<uint16_t>(parV)));
+#else
+#	error "Unsupported endianness"
+#endif
+	}
+	template <>
+	[[gnu::pure]] inline constexpr int32_t htobe (int32_t parV) {
+#if __BYTE_ORDER == __BIG_ENDIAN
+		return parV;
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+		return static_cast<int32_t>(__builtin_bswap32(static_cast<uint32_t>(parV)));
+#else
+#	error "Unsupported endianness"
+#endif
+	}
+	template <>
+	[[gnu::pure]] inline constexpr int64_t htobe (int64_t parV) {
+#if __BYTE_ORDER == __BIG_ENDIAN
+		return parV;
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+		return static_cast<int64_t>(__builtin_bswap64(static_cast<uint64_t>(parV)));
+#else
+#	error "Unsupported endianness"
+#endif
+	}
+
+	template <>
+	[[gnu::pure]] inline constexpr uint8_t htole (uint8_t parV) {
+		return parV;
+	}
+	template <>
+	[[gnu::pure]] inline constexpr uint16_t htole (uint16_t parV) {
+#if __BYTE_ORDER == __BIG_ENDIAN
+		return __builtin_bswap16(parV);
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+		return parV;
+#else
+#	error "Unsupported endianness"
+#endif
+	}
+	template <>
+	[[gnu::pure]] inline constexpr uint32_t htole (uint32_t parV) {
+#if __BYTE_ORDER == __BIG_ENDIAN
+		return __builtin_bswap32(parV);
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+		return parV;
+#else
+#	error "Unsupported endianness"
+#endif
+	}
+	template <>
+	[[gnu::pure]] inline constexpr uint64_t htole (uint64_t parV) {
+#if __BYTE_ORDER == __BIG_ENDIAN
+		return __builtin_bswap64(parV);
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+		return parV;
+#else
+#	error "Unsupported endianness"
+#endif
+	}
+	template <>
+	[[gnu::pure]] inline constexpr int8_t htole (int8_t parV) {
+		return parV;
+	}
+	template <>
+	[[gnu::pure]] inline constexpr int16_t htole (int16_t parV) {
+#if __BYTE_ORDER == __BIG_ENDIAN
+		return static_cast<int16_t>(__builtin_bswap16(static_cast<uint16_t>(parV)));
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+		return parV;
+#else
+#	error "Unsupported endianness"
+#endif
+	}
+	template <>
+	[[gnu::pure]] inline constexpr int32_t htole (int32_t parV) {
+#if __BYTE_ORDER == __BIG_ENDIAN
+		return static_cast<int32_t>(__builtin_bswap32(static_cast<uint32_t>(parV)));
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+		return parV;
+#else
+#	error "Unsupported endianness"
+#endif
+	}
+	template <>
+	[[gnu::pure]] inline constexpr int64_t htole (int64_t parV) {
+#if __BYTE_ORDER == __BIG_ENDIAN
+		return static_cast<int64_t>(__builtin_bswap64(static_cast<uint64_t>(parV)));
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+		return parV;
+#else
+#	error "Unsupported endianness"
+#endif
+	}
+
+	template <typename T>
+	[[gnu::pure]] inline constexpr T betoh (T parV) {
+		return htobe<T>(parV);
+	}
+
+	template <typename T>
+	[[gnu::pure]] inline constexpr T letoh (T parV) {
+		return htole<T>(parV);
+	}
+} //namespace dhandy
+
+#endif
diff --git a/test/unit/CMakeLists.txt b/test/unit/CMakeLists.txt
index 98e12db..4798277 100644
--- a/test/unit/CMakeLists.txt
+++ b/test/unit/CMakeLists.txt
@@ -3,6 +3,7 @@ project(dhandy_unit_test CXX)
 add_executable(${PROJECT_NAME}
 	main.cpp
 	lexical_cast_test.cpp
+	endianness_test.cpp
 )
 set_property(TARGET ${PROJECT_NAME} PROPERTY CXX_STANDARD 11)
 set_property(TARGET ${PROJECT_NAME} PROPERTY CXX_STANDARD_REQUIRED ON)
diff --git a/test/unit/endianness_test.cpp b/test/unit/endianness_test.cpp
new file mode 100644
index 0000000..4e2d584
--- /dev/null
+++ b/test/unit/endianness_test.cpp
@@ -0,0 +1,120 @@
+/* Copyright 2017, Michele Santullo
+ * This file is part of "duckhandy".
+ *
+ * "duckhandy" is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * "duckhandy" is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with "duckhandy".  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "catch.hpp"
+#include "duckhandy/endianness.hpp"
+#include <endian.h>
+
+TEST_CASE ("Host to little-endian", "[htole][endianness]") {
+	using dhandy::htole;
+#if __BYTE_ORDER == __BIG_ENDIAN
+	static_assert(htole<uint8_t>(0x12) == 0x12, "Wrong conversion");
+	static_assert(htole<uint16_t>(0x1234) == 0x3412, "Wrong conversion");
+	static_assert(htole<uint32_t>(0x12345678) == 0x78563412, "Wrong conversion");
+	static_assert(htole<uint64_t>(0x123456789ABCDEF0) == 0xF0DEBC9A78563412, "Wrong conversion");
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+	static_assert(htole<uint8_t>(0x12) == 0x12, "Wrong conversion");
+	static_assert(htole<uint16_t>(0x1234) == 0x1234, "Wrong conversion");
+	static_assert(htole<uint32_t>(0x12345678) == 0x12345678, "Wrong conversion");
+	static_assert(htole<uint64_t>(0x123456789ABCDEF0) == 0x123456789ABCDEF0, "Wrong conversion");
+#endif
+
+	CHECK(htole<uint8_t>(0x12) == 0x12);
+	CHECK(htole<uint16_t>(0x1234) == htole16(0x1234));
+	CHECK(htole<uint32_t>(0x12345678) == htole32(0x12345678));
+	CHECK(htole<uint64_t>(0x123456789ABCDEF0) == htole64(0x123456789ABCDEF0));
+
+	CHECK(htole<int8_t>(0xFE) == static_cast<int8_t>(0xFE));
+	CHECK(htole<int16_t>(0xFEDC) == static_cast<int16_t>(htole16(0xFEDC)));
+	CHECK(htole<int32_t>(0xFEDCBA09) == static_cast<int32_t>(htole32(0xFEDCBA09)));
+	CHECK(htole<int64_t>(0xFEDCBA0987654321) == static_cast<int64_t>(htole64(0xFEDCBA0987654321)));
+}
+
+TEST_CASE ("Little-endian to host", "[letoh][endianness]") {
+	using dhandy::letoh;
+#if __BYTE_ORDER == __BIG_ENDIAN
+	static_assert(letoh<uint8_t>(0x12) == 0x12, "Wrong conversion");
+	static_assert(letoh<uint16_t>(0x1234) == 0x3412, "Wrong conversion");
+	static_assert(letoh<uint32_t>(0x12345678) == 0x78563412, "Wrong conversion");
+	static_assert(letoh<uint64_t>(0x123456789ABCDEF0) == 0xF0DEBC9A78563412, "Wrong conversion");
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+	static_assert(letoh<uint8_t>(0x12) == 0x12, "Wrong conversion");
+	static_assert(letoh<uint16_t>(0x1234) == 0x1234, "Wrong conversion");
+	static_assert(letoh<uint32_t>(0x12345678) == 0x12345678, "Wrong conversion");
+	static_assert(letoh<uint64_t>(0x123456789ABCDEF0) == 0x123456789ABCDEF0, "Wrong conversion");
+#endif
+
+	CHECK(letoh<uint8_t>(0x12) == 0x12);
+	CHECK(letoh<uint16_t>(0x1234) == le16toh(0x1234));
+	CHECK(letoh<uint32_t>(0x12345678) == le32toh(0x12345678));
+	CHECK(letoh<uint64_t>(0x123456789ABCDEF0) == le64toh(0x123456789ABCDEF0));
+
+	CHECK(letoh<int8_t>(0xFE) == static_cast<int8_t>(0xFE));
+	CHECK(letoh<int16_t>(0xFEDC) == static_cast<int16_t>(le16toh(0xFEDC)));
+	CHECK(letoh<int32_t>(0xFEDCBA09) == static_cast<int32_t>(le32toh(0xFEDCBA09)));
+	CHECK(letoh<int64_t>(0xFEDCBA0987654321) == static_cast<int64_t>(le64toh(0xFEDCBA0987654321)));
+}
+
+TEST_CASE ("Host to big-endian", "[htobe][endianness]") {
+	using dhandy::htobe;
+#if __BYTE_ORDER == __BIG_ENDIAN
+	static_assert(htobe<uint8_t>(0x12) == 0x12, "Wrong conversion");
+	static_assert(htobe<uint16_t>(0x1234) == 0x1234, "Wrong conversion");
+	static_assert(htobe<uint32_t>(0x12345678) == 0x12345678, "Wrong conversion");
+	static_assert(htobe<uint64_t>(0x123456789ABCDEF0) == 0x123456789ABCDEF0, "Wrong conversion");
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+	static_assert(htobe<uint8_t>(0x12) == 0x12, "Wrong conversion");
+	static_assert(htobe<uint16_t>(0x1234) == 0x3412, "Wrong conversion");
+	static_assert(htobe<uint32_t>(0x12345678) == 0x78563412, "Wrong conversion");
+	static_assert(htobe<uint64_t>(0x123456789ABCDEF0) == 0xF0DEBC9A78563412, "Wrong conversion");
+#endif
+
+	CHECK(htobe<uint8_t>(0x12) == 0x12);
+	CHECK(htobe<uint16_t>(0x1234) == htobe16(0x1234));
+	CHECK(htobe<uint32_t>(0x12345678) == htobe32(0x12345678));
+	CHECK(htobe<uint64_t>(0x123456789ABCDEF0) == htobe64(0x123456789ABCDEF0));
+
+	CHECK(htobe<int8_t>(0xFE) == static_cast<int8_t>(0xFE));
+	CHECK(htobe<int16_t>(0xFEDC) == static_cast<int16_t>(htobe16(0xFEDC)));
+	CHECK(htobe<int32_t>(0xFEDCBA09) == static_cast<int32_t>(htobe32(0xFEDCBA09)));
+	CHECK(htobe<int64_t>(0xFEDCBA0987654321) == static_cast<int64_t>(htobe64(0xFEDCBA0987654321)));
+}
+
+TEST_CASE ("Big-endian to host", "[betoh][endianness]") {
+	using dhandy::betoh;
+#if __BYTE_ORDER == __BIG_ENDIAN
+	static_assert(betoh<uint8_t>(0x12) == 0x12, "Wrong conversion");
+	static_assert(betoh<uint16_t>(0x1234) == 0x1234, "Wrong conversion");
+	static_assert(betoh<uint32_t>(0x12345678) == 0x12345678, "Wrong conversion");
+	static_assert(betoh<uint64_t>(0x123456789ABCDEF0) == 0x123456789ABCDEF0, "Wrong conversion");
+#elif __BYTE_ORDER == __LITTLE_ENDIAN
+	static_assert(betoh<uint8_t>(0x12) == 0x12, "Wrong conversion");
+	static_assert(betoh<uint16_t>(0x1234) == 0x3412, "Wrong conversion");
+	static_assert(betoh<uint32_t>(0x12345678) == 0x78563412, "Wrong conversion");
+	static_assert(betoh<uint64_t>(0x123456789ABCDEF0) == 0xF0DEBC9A78563412, "Wrong conversion");
+#endif
+
+	CHECK(betoh<uint8_t>(0x12) == 0x12);
+	CHECK(betoh<uint16_t>(0x1234) == be16toh(0x1234));
+	CHECK(betoh<uint32_t>(0x12345678) == be32toh(0x12345678));
+	CHECK(betoh<uint64_t>(0x123456789ABCDEF0) == be64toh(0x123456789ABCDEF0));
+
+	CHECK(betoh<int8_t>(0xFE) == static_cast<int8_t>(0xFE));
+	CHECK(betoh<int16_t>(0xFEDC) == static_cast<int16_t>(be16toh(0xFEDC)));
+	CHECK(betoh<int32_t>(0xFEDCBA09) == static_cast<int32_t>(be32toh(0xFEDCBA09)));
+	CHECK(betoh<int64_t>(0xFEDCBA0987654321) == static_cast<int64_t>(be64toh(0xFEDCBA0987654321)));
+}