diff --git a/libluna/CMakeLists.txt b/libluna/CMakeLists.txt
index 7c110a14..a775b140 100644
--- a/libluna/CMakeLists.txt
+++ b/libluna/CMakeLists.txt
@@ -23,6 +23,7 @@ set(FREESTANDING_SOURCES
     src/Spinlock.cpp
     src/PathParser.cpp
     src/UBSAN.cpp
+    src/Base64.cpp
 )
 
 set(SOURCES
diff --git a/libluna/include/luna/Base64.h b/libluna/include/luna/Base64.h
new file mode 100644
index 00000000..3efe79b8
--- /dev/null
+++ b/libluna/include/luna/Base64.h
@@ -0,0 +1,13 @@
+#pragma once
+#include <luna/Buffer.h>
+#include <luna/String.h>
+
+namespace Base64
+{
+    Result<String> encode(StringView data);
+    Result<String> encode(Slice<const u8> data);
+    Result<String> encode(const Buffer& data);
+
+    Result<Buffer> decode(StringView data, bool allow_garbage_chars = false);
+    Result<String> decode_string(StringView data, bool allow_garbage_chars = false);
+}
diff --git a/libluna/src/Base64.cpp b/libluna/src/Base64.cpp
new file mode 100644
index 00000000..7af1b0e5
--- /dev/null
+++ b/libluna/src/Base64.cpp
@@ -0,0 +1,144 @@
+#include <luna/Base64.h>
+#include <luna/CType.h>
+#include <luna/DebugLog.h>
+#include <luna/Slice.h>
+#include <luna/StringBuilder.h>
+
+static const char g_base64_alphabet[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
+
+static Option<u8> value_from_base64_character(char letter)
+{
+    if (_isupper(letter)) return (u8)(letter - 'A');
+    if (_islower(letter)) return (u8)(26 + (letter - 'a'));
+    if (_isdigit(letter)) return (u8)(52 + (letter - '0'));
+    if (letter == '+') return 62;
+    if (letter == '/') return 63;
+    if (letter == '=') return {};
+    unreachable();
+}
+
+static Result<void> decode_base64_buffer(const Vector<char>& input, Buffer& output)
+{
+    const usize bytes_to_decode = 4;
+
+    Option<u8> base64_data[bytes_to_decode];
+
+    for (usize i = 0; i < 4; i++) { base64_data[i] = value_from_base64_character(input[i]); }
+
+    u8 decoded_bytes[3];
+    usize total_decoded_bytes = 0;
+
+    decoded_bytes[0] = (base64_data[0].value() << 2) | (base64_data[1].value() >> 4);
+    total_decoded_bytes++;
+
+    if (base64_data[2].has_value())
+    {
+        decoded_bytes[1] = ((base64_data[1].value() & 0b1111) << 4) | (base64_data[2].value() >> 2);
+        total_decoded_bytes++;
+    }
+
+    if (base64_data[3].has_value())
+    {
+        decoded_bytes[2] = ((base64_data[2].value() & 0b11) << 6) | base64_data[3].value();
+        total_decoded_bytes++;
+    }
+
+    return output.append_data(decoded_bytes, total_decoded_bytes);
+}
+
+namespace Base64
+{
+    Result<String> encode(StringView data)
+    {
+        return encode(Slice<const u8> { (const u8*)data.chars(), data.length() });
+    }
+
+    Result<String> encode(const Buffer& data)
+    {
+        return encode(Slice<const u8> { data.data(), data.size() });
+    }
+
+    Result<String> encode(Slice<const u8> data)
+    {
+        StringBuilder sb;
+
+        for (usize i = 0; i < data.size(); i += 3)
+        {
+            usize bytes_to_encode = data.size() - i;
+            if (bytes_to_encode > 3) bytes_to_encode = 3;
+
+            TRY(sb.add(g_base64_alphabet[data[i] >> 2]));
+            if (bytes_to_encode > 1) TRY(sb.add(g_base64_alphabet[((data[i] & 0b11) << 4) | (data[i + 1] >> 4)]));
+            else
+            {
+                TRY(sb.add(g_base64_alphabet[(data[i] & 0b11) << 6]));
+                TRY(sb.add("=="_sv));
+                break;
+            }
+
+            if (bytes_to_encode > 2) TRY(sb.add(g_base64_alphabet[((data[i + 1] & 0b1111) << 2) | (data[i + 2] >> 6)]));
+            else
+            {
+                TRY(sb.add(g_base64_alphabet[(data[i + 1] & 0b1111) << 2]));
+                TRY(sb.add("="_sv));
+                break;
+            }
+
+            TRY(sb.add(g_base64_alphabet[data[i + 2] & 0b111111]));
+        }
+
+        return sb.string();
+    }
+
+    Result<Buffer> decode(StringView data, bool allow_garbage_chars)
+    {
+        Buffer buf;
+
+        char* padding = strchr(data.chars(), '=');
+        if (padding)
+        {
+            // If the string ends with padding, it must be either one or two equals signs.
+            if (padding[1] != '=' && padding[1] != '\0') return err(EINVAL);
+            if (padding[1])
+            {
+                if (padding[strspn(&padding[2], "\n") + 2]) return err(EINVAL);
+            }
+        }
+
+        Vector<char> chars_read;
+
+        for (const auto& c : data)
+        {
+            if (c == '\n') continue;
+
+            if (!_isalnum(c) && c != '+' && c != '/' && c != '=')
+            {
+                if (allow_garbage_chars) continue;
+                return err(EINVAL);
+            }
+
+            TRY(chars_read.try_append(c));
+
+            if (chars_read.size() == 4)
+            {
+                TRY(decode_base64_buffer(chars_read, buf));
+                chars_read.clear();
+            }
+        }
+
+        // Unterminated input
+        if (chars_read.size() > 0) return err(EINVAL);
+
+        return buf;
+    }
+
+    Result<String> decode_string(StringView data, bool allow_garbage_chars)
+    {
+        auto buf = TRY(decode(data, allow_garbage_chars));
+
+        u8 nul_byte = '\0';
+        TRY(buf.append_data(&nul_byte, 1));
+
+        return String::from_cstring((char*)buf.data());
+    }
+}