diff --git a/src/openrct2/CMakeLists.txt b/src/openrct2/CMakeLists.txt
index a141e10d96..8dcd0234d2 100644
--- a/src/openrct2/CMakeLists.txt
+++ b/src/openrct2/CMakeLists.txt
@@ -196,8 +196,6 @@ if (HAVE_DISCORD_RPC)
     target_link_libraries(libopenrct2 discord-rpc)
 endif()
 
-target_link_libraries(libopenrct2 fribidi)
-
 # Includes
 target_include_directories(${PROJECT_NAME} PRIVATE ${LIBZIP_INCLUDE_DIRS})
 target_include_directories(${PROJECT_NAME} PRIVATE ${PNG_INCLUDE_DIRS}
diff --git a/src/openrct2/localisation/LanguagePack.cpp b/src/openrct2/localisation/LanguagePack.cpp
index b24ba013bc..81f19531e4 100644
--- a/src/openrct2/localisation/LanguagePack.cpp
+++ b/src/openrct2/localisation/LanguagePack.cpp
@@ -9,13 +9,6 @@
 
 #include "LanguagePack.h"
 
-extern "C" {
-#include <fribidi/fribidi-bidi-types.h>
-#include <fribidi/fribidi-char-sets.h>
-#include <fribidi/fribidi-flags.h>
-#include <fribidi/fribidi.h>
-}
-
 #include "../common.h"
 #include "../core/FileStream.hpp"
 #include "../core/Memory.hpp"
@@ -29,6 +22,22 @@ extern "C" {
 #include <string>
 #include <vector>
 
+#ifdef _WIN32
+extern "C" {
+#    include <fribidi/fribidi-bidi-types.h>
+#    include <fribidi/fribidi-char-sets.h>
+#    include <fribidi/fribidi-flags.h>
+#    include <fribidi/fribidi.h>
+}
+#else
+#    include <unicode/ubidi.h>
+#    include <unicode/unistr.h>
+#    include <unicode/ushape.h>
+#    include <unicode/ustring.h>
+#    include <unicode/utf.h>
+#    include <unicode/utypes.h>
+#endif
+
 // Don't try to load more than language files that exceed 64 MiB
 constexpr uint64_t MAX_LANGUAGE_SIZE = 64 * 1024 * 1024;
 constexpr uint64_t MAX_OBJECT_OVERRIDES = 4096;
@@ -649,6 +658,7 @@ private:
 
     std::string FixRTL(std::string& input)
     {
+#ifdef _WIN32
         FriBidiChar utf32String[1024] = { 0 };
         FriBidiStrIndex len = input.length() + 1;
         fribidi_charset_to_unicode(FRIBIDI_CHAR_SET_UTF8, input.c_str(), len, utf32String);
@@ -672,6 +682,31 @@ private:
         fribidi_unicode_to_charset(FRIBIDI_CHAR_SET_UTF8, reorderedStr, len, outputString);
 
         return std::string(outputString);
+#else
+        UErrorCode err = static_cast<UErrorCode>(0);
+        // Force a hard left-to-right at the beginning (will mess up mixed strings' word order otherwise)
+        std::string text2 = std::string(u8"\xE2\x80\xAA") + input;
+
+        icu::UnicodeString ustr = icu::UnicodeString::fromUTF8(icu::StringPiece(text2));
+
+        int32_t length = ustr.length();
+        icu::UnicodeString reordered;
+        icu::UnicodeString shaped;
+        UBiDi* bidi = ubidi_openSized(length, 0, &err);
+        // UBIDI_DEFAULT_LTR preserves formatting codes.
+        ubidi_setPara(bidi, ustr.getBuffer(), length, UBIDI_DEFAULT_LTR, nullptr, &err);
+        ubidi_writeReordered(bidi, reordered.getBuffer(length), length, UBIDI_DO_MIRRORING | UBIDI_REMOVE_BIDI_CONTROLS, &err);
+        ubidi_close(bidi);
+        reordered.releaseBuffer(length);
+        u_shapeArabic(
+            reordered.getBuffer(), length, shaped.getBuffer(length), length,
+            U_SHAPE_LETTERS_SHAPE | U_SHAPE_LENGTH_FIXED_SPACES_NEAR | U_SHAPE_TEXT_DIRECTION_VISUAL_LTR, &err);
+        shaped.releaseBuffer(length);
+
+        std::string cppstring;
+        shaped.toUTF8String(cppstring);
+        return cppstring;
+#endif
     }
 };