diff --git a/libc/src/__support/wchar/CMakeLists.txt b/libc/src/__support/wchar/CMakeLists.txt
index d3fb58ed0c71c..0828e4057c172 100644
--- a/libc/src/__support/wchar/CMakeLists.txt
+++ b/libc/src/__support/wchar/CMakeLists.txt
@@ -6,6 +6,19 @@ add_header_library(
     libc.hdr.types.char32_t    
 )
 
+add_header_library(
+  string_converter
+  HDRS
+    string_converter.h
+  DEPENDS
+    libc.hdr.types.char8_t
+    libc.hdr.types.char32_t
+    libc.hdr.types.size_t
+    libc.src.__support.error_or
+    .mbstate
+    .character_converter 
+)
+
 add_object_library(
   character_converter
   HDRS
diff --git a/libc/src/__support/wchar/character_converter.cpp b/libc/src/__support/wchar/character_converter.cpp
index 3cacfa5689e4d..8d7c4183c5998 100644
--- a/libc/src/__support/wchar/character_converter.cpp
+++ b/libc/src/__support/wchar/character_converter.cpp
@@ -14,6 +14,7 @@
 #include "src/__support/error_or.h"
 #include "src/__support/math_extras.h"
 #include "src/__support/wchar/mbstate.h"
+#include <stddef.h>
 
 #include "character_converter.h"
 
@@ -92,6 +93,7 @@ int CharacterConverter::push(char8_t utf8_byte) {
     state->bytes_stored++;
     return 0;
   }
+
   // Invalid byte -> reset the state
   clear();
   return EILSEQ;
@@ -130,6 +132,12 @@ ErrorOr<char32_t> CharacterConverter::pop_utf32() {
   return utf32;
 }
 
+size_t CharacterConverter::sizeAsUTF32() {
+  return 1; // a single utf-32 value can fit an entire character
+}
+
+size_t CharacterConverter::sizeAsUTF8() { return state->total_bytes; }
+
 ErrorOr<char8_t> CharacterConverter::pop_utf8() {
   if (isEmpty())
     return Error(-1);
@@ -156,6 +164,9 @@ ErrorOr<char8_t> CharacterConverter::pop_utf8() {
   }
 
   state->bytes_stored--;
+  if (state->bytes_stored == 0)
+    clear();
+
   return static_cast<char8_t>(output);
 }
 
diff --git a/libc/src/__support/wchar/character_converter.h b/libc/src/__support/wchar/character_converter.h
index d9a63fdc0522c..9e8dd71028002 100644
--- a/libc/src/__support/wchar/character_converter.h
+++ b/libc/src/__support/wchar/character_converter.h
@@ -14,6 +14,7 @@
 #include "src/__support/common.h"
 #include "src/__support/error_or.h"
 #include "src/__support/wchar/mbstate.h"
+#include <stddef.h>
 
 namespace LIBC_NAMESPACE_DECL {
 namespace internal {
@@ -30,6 +31,9 @@ class CharacterConverter {
   bool isEmpty();
   bool isValidState();
 
+  size_t sizeAsUTF32();
+  size_t sizeAsUTF8();
+
   int push(char8_t utf8_byte);
   int push(char32_t utf32);
 
diff --git a/libc/src/__support/wchar/string_converter.h b/libc/src/__support/wchar/string_converter.h
new file mode 100644
index 0000000000000..8c0417a3b7df4
--- /dev/null
+++ b/libc/src/__support/wchar/string_converter.h
@@ -0,0 +1,113 @@
+//===-- Definition of a class for mbstate_t and conversion -----*-- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_LIBC_SRC___SUPPORT_STRING_CONVERTER_H
+#define LLVM_LIBC_SRC___SUPPORT_STRING_CONVERTER_H
+
+#include "hdr/types/char32_t.h"
+#include "hdr/types/char8_t.h"
+#include "hdr/types/size_t.h"
+#include "src/__support/common.h"
+#include "src/__support/error_or.h"
+#include "src/__support/wchar/character_converter.h"
+#include "src/__support/wchar/mbstate.h"
+
+namespace LIBC_NAMESPACE_DECL {
+namespace internal {
+
+template <typename T> class StringConverter {
+private:
+  CharacterConverter cr;
+  const T *src;
+  size_t src_len;
+  size_t src_idx;
+  size_t num_pushed;
+  size_t num_to_write;
+
+  int pushFullCharacter() {
+    for (num_pushed = 0; !cr.isFull() && src_idx + num_pushed < src_len;
+         ++num_pushed) {
+      int err = cr.push(src[src_idx + num_pushed]);
+      if (err != 0)
+        return err;
+    }
+
+    // if we aren't able to read a full character from the source string
+    if (src_idx + num_pushed == src_len && !cr.isFull()) {
+      src_idx += num_pushed;
+      return -1;
+    }
+
+    return 0;
+  }
+
+public:
+  StringConverter(const T *s, size_t srclen, size_t dstlen, mbstate *ps)
+      : cr(ps), src(s), src_len(srclen), src_idx(0), num_pushed(0),
+        num_to_write(dstlen) {
+    pushFullCharacter();
+  }
+
+  StringConverter(const T *s, size_t dstlen, mbstate *ps)
+      : StringConverter(s, SIZE_MAX, dstlen, ps) {}
+
+  ErrorOr<char32_t> popUTF32() {
+    if (cr.isEmpty()) {
+      int err = pushFullCharacter();
+      if (err != 0)
+        return Error(err);
+
+      if (cr.sizeAsUTF32() > num_to_write) {
+        cr.clear();
+        return Error(-1);
+      }
+    }
+
+    auto out = cr.pop_utf32();
+    if (cr.isEmpty())
+      src_idx += num_pushed;
+
+    if (out.has_value() && out.value() == L'\0')
+      src_len = src_idx;
+
+    num_to_write--;
+
+    return out;
+  }
+
+  ErrorOr<char8_t> popUTF8() {
+    if (cr.isEmpty()) {
+      int err = pushFullCharacter();
+      if (err != 0)
+        return Error(err);
+
+      if (cr.sizeAsUTF8() > num_to_write) {
+        cr.clear();
+        return Error(-1);
+      }
+    }
+
+    auto out = cr.pop_utf8();
+    if (cr.isEmpty())
+      src_idx += num_pushed;
+
+    if (out.has_value() && out.value() == '\0')
+      src_len = src_idx;
+
+    num_to_write--;
+
+    return out;
+  }
+
+  size_t getSourceIndex() { return src_idx; }
+};
+
+} // namespace internal
+} // namespace LIBC_NAMESPACE_DECL
+
+#endif // LLVM_LIBC_SRC___SUPPORT_STRING_CONVERTER_H
diff --git a/libc/test/src/__support/wchar/CMakeLists.txt b/libc/test/src/__support/wchar/CMakeLists.txt
index 5176bfd4b024b..6982232d67544 100644
--- a/libc/test/src/__support/wchar/CMakeLists.txt
+++ b/libc/test/src/__support/wchar/CMakeLists.txt
@@ -19,3 +19,17 @@ add_libc_test(
   DEPENDS
     libc.src.__support.wchar.character_converter
 )
+
+add_libc_test(
+  string_converter_test.cpp
+  SUITE
+    libc-support-tests
+  SRCS
+    string_converter_test.cpp
+  DEPENDS
+    libc.src.__support.wchar.string_converter
+    libc.src.__support.wchar.mbstate
+    libc.src.__support.error_or
+    libc.hdr.errno_macros
+    libc.hdr.types.char32_t
+)
diff --git a/libc/test/src/__support/wchar/string_converter_test.cpp b/libc/test/src/__support/wchar/string_converter_test.cpp
new file mode 100644
index 0000000000000..cb908e2fec2a8
--- /dev/null
+++ b/libc/test/src/__support/wchar/string_converter_test.cpp
@@ -0,0 +1,308 @@
+//===-- Unittests for StringConverter class -------------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "hdr/errno_macros.h"
+#include "hdr/types/char32_t.h"
+#include "hdr/types/char8_t.h"
+#include "src/__support/error_or.h"
+#include "src/__support/wchar/mbstate.h"
+#include "src/__support/wchar/string_converter.h"
+#include "test/UnitTest/Test.h"
+
+TEST(LlvmLibcStringConverterTest, UTF8To32) {
+  // first 4 bytes are clown emoji (🤡), then next 3 are sigma symbol (∑)
+  const char *src = "\xF0\x9F\xA4\xA1\xE2\x88\x91";
+  LIBC_NAMESPACE::internal::mbstate state;
+  LIBC_NAMESPACE::internal::StringConverter<char8_t> sc(
+      reinterpret_cast<const char8_t *>(src), SIZE_MAX, &state);
+
+  auto res = sc.popUTF32();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0x1f921);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 4);
+
+  res = sc.popUTF32();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0x2211);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 7);
+
+  res = sc.popUTF32();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 8);
+
+  res = sc.popUTF32();
+  ASSERT_FALSE(res.has_value());
+  ASSERT_EQ(res.error(), -1);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 8);
+}
+
+TEST(LlvmLibcStringConverterTest, UTF32To8) {
+  const wchar_t *src = L"\x1f921\x2211"; // clown emoji, sigma symbol
+  LIBC_NAMESPACE::internal::mbstate state;
+  LIBC_NAMESPACE::internal::StringConverter<char32_t> sc(
+      reinterpret_cast<const char32_t *>(src), SIZE_MAX, &state);
+
+  auto res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0xF0);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 0);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0x9F);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 0);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0xA4);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 0);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0xA1);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 1);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0xE2);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 1);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0x88);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 1);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0x91);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 2);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 3);
+
+  res = sc.popUTF8();
+  ASSERT_FALSE(res.has_value());
+  ASSERT_EQ(res.error(), -1);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 3);
+}
+
+TEST(LlvmLibcStringConverterTest, UTF32To8PartialRead) {
+  const wchar_t *src = L"\x1f921\x2211"; // clown emoji, sigma symbol
+  LIBC_NAMESPACE::internal::mbstate state;
+  LIBC_NAMESPACE::internal::StringConverter<char32_t> sc(
+      reinterpret_cast<const char32_t *>(src), 1, SIZE_MAX, &state);
+
+  auto res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0xF0);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 0);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0x9F);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 0);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0xA4);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 0);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0xA1);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 1);
+
+  res = sc.popUTF8();
+  ASSERT_FALSE(res.has_value());
+  ASSERT_EQ(res.error(), -1);
+}
+
+TEST(LlvmLibcStringConverterTest, UTF8To32PartialRead) {
+  // first 4 bytes are clown emoji, then next 3 are sigma symbol
+  const char *src = "\xF0\x9F\xA4\xA1\xE2\x88\x91";
+  LIBC_NAMESPACE::internal::mbstate state;
+  LIBC_NAMESPACE::internal::StringConverter<char8_t> sc(
+      reinterpret_cast<const char8_t *>(src), 5, SIZE_MAX, &state);
+
+  auto res = sc.popUTF32();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0x1f921);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 4);
+
+  res = sc.popUTF32();
+  ASSERT_FALSE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.error()), -1);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 5);
+}
+
+TEST(LlvmLibcStringConverterTest, UTF32To8ErrorHandling) {
+  const wchar_t *src = L"\x1f921\xffffff"; // clown emoji, invalid utf32
+  LIBC_NAMESPACE::internal::mbstate state;
+  LIBC_NAMESPACE::internal::StringConverter<char32_t> sc(
+      reinterpret_cast<const char32_t *>(src), SIZE_MAX, &state);
+
+  auto res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0xF0);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 0);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0x9F);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 0);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0xA4);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 0);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0xA1);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 1);
+
+  res = sc.popUTF8();
+  ASSERT_FALSE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.error()), EILSEQ);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 1);
+}
+
+TEST(LlvmLibcStringConverterTest, UTF8To32ErrorHandling) {
+  // first 4 bytes are clown emoji (🤡)
+  // next 3 form an invalid character
+  const char *src = "\xF0\x9F\xA4\xA1\x90\x88\x30";
+  LIBC_NAMESPACE::internal::mbstate state;
+  LIBC_NAMESPACE::internal::StringConverter<char8_t> sc(
+      reinterpret_cast<const char8_t *>(src), SIZE_MAX, &state);
+
+  auto res = sc.popUTF32();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0x1f921);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 4);
+
+  res = sc.popUTF32();
+  ASSERT_FALSE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.error()), EILSEQ);
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 4);
+}
+
+TEST(LlvmLibcStringConverterTest, MultipleStringConverters32To8) {
+  /*
+  We do NOT test partially popping a character and expecting the next
+  StringConverter to continue where we left off. This is not expected to work
+  and considered invalid.
+  */
+  const wchar_t *src = L"\x1f921\xff"; // clown emoji, sigma symbol
+  LIBC_NAMESPACE::internal::mbstate state;
+  LIBC_NAMESPACE::internal::StringConverter<char32_t> sc1(
+      reinterpret_cast<const char32_t *>(src), 1, SIZE_MAX, &state);
+
+  auto res = sc1.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0xF0);
+  ASSERT_EQ(static_cast<int>(sc1.getSourceIndex()), 0);
+
+  res = sc1.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0x9F);
+  ASSERT_EQ(static_cast<int>(sc1.getSourceIndex()), 0);
+
+  res = sc1.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0xA4);
+  ASSERT_EQ(static_cast<int>(sc1.getSourceIndex()), 0);
+
+  res = sc1.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0xA1);
+  ASSERT_EQ(static_cast<int>(sc1.getSourceIndex()), 1);
+
+  LIBC_NAMESPACE::internal::StringConverter<char32_t> sc2(
+      reinterpret_cast<const char32_t *>(src) + sc1.getSourceIndex(), 1,
+      SIZE_MAX, &state);
+
+  res = sc2.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0xC3);
+  ASSERT_EQ(static_cast<int>(sc2.getSourceIndex()), 0);
+
+  res = sc2.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0xBF);
+  ASSERT_EQ(static_cast<int>(sc2.getSourceIndex()), 1);
+}
+
+TEST(LlvmLibcStringConverterTest, MultipleStringConverters8To32) {
+  const char *src = "\xF0\x9F\xA4\xA1"; // clown emoji
+  LIBC_NAMESPACE::internal::mbstate state;
+  LIBC_NAMESPACE::internal::StringConverter<char8_t> sc1(
+      reinterpret_cast<const char8_t *>(src), 2, SIZE_MAX, &state);
+
+  auto res = sc1.popUTF32();
+  ASSERT_FALSE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.error()), -1);
+  ASSERT_EQ(static_cast<int>(sc1.getSourceIndex()), 2);
+
+  LIBC_NAMESPACE::internal::StringConverter<char8_t> sc2(
+      reinterpret_cast<const char8_t *>(src) + sc1.getSourceIndex(), 3,
+      SIZE_MAX, &state);
+
+  res = sc2.popUTF32();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0x1f921);
+  ASSERT_EQ(static_cast<int>(sc2.getSourceIndex()), 2);
+
+  res = sc2.popUTF32();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(res.value()), 0);
+  ASSERT_EQ(static_cast<int>(sc2.getSourceIndex()), 3);
+}
+
+TEST(LlvmLibcStringConverterTest, DstLimitUTF8To32) {
+  const char *src = "\xF0\x9F\xA4\xA1\xF0\x9F\xA4\xA1"; // 2 clown emojis
+  LIBC_NAMESPACE::internal::mbstate state;
+  LIBC_NAMESPACE::internal::StringConverter<char8_t> sc(
+      reinterpret_cast<const char8_t *>(src), SIZE_MAX, 1, &state);
+
+  auto res = sc.popUTF32();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 4);
+
+  res = sc.popUTF32(); // no space to pop this into
+  ASSERT_FALSE(res.has_value());
+}
+
+TEST(LlvmLibcStringConverterTest, DstLimitUTF32To8) {
+  const wchar_t *src = L"\x1f921\x1f921"; // 2 clown emojis
+  LIBC_NAMESPACE::internal::mbstate state;
+  LIBC_NAMESPACE::internal::StringConverter<char32_t> sc(
+      reinterpret_cast<const char32_t *>(src), SIZE_MAX, 5, &state);
+
+  auto res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 0);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 0);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 0);
+
+  res = sc.popUTF8();
+  ASSERT_TRUE(res.has_value());
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 1);
+
+  res = sc.popUTF8();
+  ASSERT_FALSE(res.has_value());
+  ASSERT_EQ(static_cast<int>(sc.getSourceIndex()), 1);
+}