fix: support non-ASCII characters (Cyrillic, CJK, emoji) in humanized typing

cloakbrowser/human/keyboard.py

@@ -41,7 +41,15 @@ def _get_nearby_key(ch: str) -> str:
 
 def human_type(page: Any, raw: RawKeyboard, text: str, cfg: HumanConfig) -> None:
     for i, ch in enumerate(text):
-        # Mistype chance — press wrong key, notice, backspace, then correct
+        # Non-ASCII characters (Cyrillic, CJK, emoji) — use insertText
+        if not ch.isascii():
+            sleep_ms(rand_range(cfg.key_hold))
+            raw.insert_text(ch)
+            if i < len(text) - 1:
+                _inter_char_delay(cfg)
+            continue
+
+        # Mistype chance — only for ASCII alphanumeric
         if random.random() < cfg.mistype_chance and ch.isalnum():
             wrong = _get_nearby_key(ch)
             _type_normal_char(raw, wrong, cfg)

cloakbrowser/human/keyboard_async.py

@@ -22,7 +22,15 @@ class AsyncRawKeyboard(Protocol):
 
 async def async_human_type(page: Any, raw: AsyncRawKeyboard, text: str, cfg: HumanConfig) -> None:
     for i, ch in enumerate(text):
-        # Mistype chance — press wrong key, notice, backspace, then correct
+        # Non-ASCII characters (Cyrillic, CJK, emoji) — use insertText
+        if not ch.isascii():
+            await async_sleep_ms(rand_range(cfg.key_hold))
+            await raw.insert_text(ch)
+            if i < len(text) - 1:
+                await _inter_char_delay(cfg)
+            continue
+
+        # Mistype chance — only for ASCII alphanumeric
         if random.random() < cfg.mistype_chance and ch.isalnum():
             wrong = _get_nearby_key(ch)
             await _type_normal_char(raw, wrong, cfg)

js/src/human/keyboard.ts

@@ -22,6 +22,11 @@ const NEARBY_KEYS: Record<string, string> = {
   '6': '57ty', '7': '68yu', '8': '79ui', '9': '80io', '0': '9p',
 };
 
+function isAscii(ch: string): boolean {
+  const code = ch.codePointAt(0);
+  return code !== undefined && code < 128;
+}
+
 function getNearbyKey(ch: string): string {
   const lower = ch.toLowerCase();
   if (lower in NEARBY_KEYS) {
@@ -38,11 +43,23 @@ export async function humanType(
   text: string,
   cfg: HumanConfig,
 ): Promise<void> {
-  for (let i = 0; i < text.length; i++) {
+  const chars = [...text]; // Handle emoji surrogate pairs correctly
-    const ch = text[i];
 
-    // Mistype chance — press wrong key, notice, backspace, then correct
+  for (let i = 0; i < chars.length; i++) {
-    if (Math.random() < cfg.mistype_chance && /[a-zA-Z0-9]/.test(ch)) {
+    const ch = chars[i];
+
+    // Non-ASCII characters (Cyrillic, CJK, emoji) — use insertText
+    if (!isAscii(ch)) {
+      await sleep(randRange(cfg.key_hold));
+      await raw.insertText(ch);
+      if (i < chars.length - 1) {
+        await interCharDelay(cfg);
+      }
+      continue;
+    }
+
+    // Mistype chance — only for ASCII alphanumeric
+    if (Math.random() < cfg.mistype_chance && /^[a-zA-Z0-9]$/.test(ch)) {
       const wrong = getNearbyKey(ch);
       await typeNormalChar(raw, wrong, cfg);
       await sleep(randRange(cfg.mistype_delay_notice));
@@ -60,7 +77,7 @@ export async function humanType(
       await typeNormalChar(raw, ch, cfg);
     }
 
-    if (i < text.length - 1) {
+    if (i < chars.length - 1) {
       await interCharDelay(cfg);
     }
   }

js/tests/humanize.test.ts

@@ -530,6 +530,81 @@ function buildMockPage(overrides: Record<string, any> = {}): any {
   return page;
 }
 
+// =========================================================================
+// humanType non-ASCII
+// =========================================================================
+describe("humanType non-ASCII", () => {
+  function makeRawKeyboardMock() {
+    const downKeys: string[] = [];
+    const insertedChars: string[] = [];
+    const raw = {
+      down: vi.fn(async (k: string) => { downKeys.push(k); }),
+      up: vi.fn(async () => {}),
+      type: vi.fn(async () => {}),
+      insertText: vi.fn(async (t: string) => { insertedChars.push(t); }),
+    };
+    return { raw, downKeys, insertedChars };
+  }
+
+  it("types Cyrillic via insertText, not down", async () => {
+    const { humanType } = await import("../src/human/keyboard.js");
+    const cfg = resolveConfig("default", { mistype_chance: 0 });
+    const { raw, downKeys, insertedChars } = makeRawKeyboardMock();
+
+    await humanType({} as any, raw, "Привет", cfg);
+
+    expect(insertedChars.join("")).toBe("Привет");
+    for (const k of downKeys) {
+      expect(k.charCodeAt(0)).toBeLessThan(128);
+    }
+  });
+
+  it("types mixed ASCII + Cyrillic correctly", async () => {
+    const { humanType } = await import("../src/human/keyboard.js");
+    const cfg = resolveConfig("default", { mistype_chance: 0 });
+    const { raw, downKeys, insertedChars } = makeRawKeyboardMock();
+
+    await humanType({} as any, raw, "Hi Мир", cfg);
+
+    expect(downKeys).toContain("H");
+    expect(downKeys).toContain("i");
+    expect(insertedChars.join("")).toContain("М");
+    expect(insertedChars.join("")).toContain("и");
+    expect(insertedChars.join("")).toContain("р");
+  });
+
+  it("types CJK via insertText", async () => {
+    const { humanType } = await import("../src/human/keyboard.js");
+    const cfg = resolveConfig("default", { mistype_chance: 0 });
+    const { raw, insertedChars } = makeRawKeyboardMock();
+
+    await humanType({} as any, raw, "你好", cfg);
+
+    expect(insertedChars.join("")).toBe("你好");
+  });
+
+  it("types emoji via insertText", async () => {
+    const { humanType } = await import("../src/human/keyboard.js");
+    const cfg = resolveConfig("default", { mistype_chance: 0 });
+    const { raw, insertedChars } = makeRawKeyboardMock();
+
+    await humanType({} as any, raw, "Hi 👋", cfg);
+
+    expect(insertedChars.join("")).toContain("👋");
+  });
+
+  it("mistype only triggers for ASCII, not Cyrillic", async () => {
+    const { humanType } = await import("../src/human/keyboard.js");
+    const cfg = resolveConfig("default", { mistype_chance: 1.0 });
+    const { raw, downKeys } = makeRawKeyboardMock();
+
+    await humanType({} as any, raw, "AБ", cfg);
+
+    expect(downKeys).toContain("Backspace");
+  });
+});
+
+
 
 function buildMockFrame(): any {
   return {

tests/test_humanize_unit.py

@@ -406,6 +406,127 @@ class TestSelectAllPlatform:
             assert _SELECT_ALL == "Control+a"
 
 
+# =========================================================================
+# 11. Non-ASCII keyboard input
+# =========================================================================
+
+class TestNonAsciiKeyboard:
+    def test_cyrillic_uses_insert_text(self):
+        from cloakbrowser.human.keyboard import human_type
+        from cloakbrowser.human.config import resolve_config
+        from unittest.mock import MagicMock
+
+        cfg = resolve_config("default", {"mistype_chance": 0})
+        page = MagicMock()
+        raw = MagicMock()
+
+        down_keys = []
+        inserted = []
+        raw.down = MagicMock(side_effect=lambda k: down_keys.append(k))
+        raw.up = MagicMock()
+        raw.insert_text = MagicMock(side_effect=lambda t: inserted.append(t))
+
+        human_type(page, raw, "Привет", cfg)
+
+        assert "".join(inserted) == "Привет"
+        for k in down_keys:
+            assert ord(k[0]) < 128 or k in ("Shift", "Backspace")
+
+    def test_mixed_ascii_cyrillic(self):
+        from cloakbrowser.human.keyboard import human_type
+        from cloakbrowser.human.config import resolve_config
+        from unittest.mock import MagicMock
+
+        cfg = resolve_config("default", {"mistype_chance": 0})
+        page = MagicMock()
+        raw = MagicMock()
+
+        down_keys = []
+        inserted = []
+        raw.down = MagicMock(side_effect=lambda k: down_keys.append(k))
+        raw.up = MagicMock()
+        raw.insert_text = MagicMock(side_effect=lambda t: inserted.append(t))
+
+        human_type(page, raw, "Hi Мир", cfg)
+
+        assert "H" in down_keys
+        assert "i" in down_keys
+        assert "М" in "".join(inserted)
+
+    def test_cjk_uses_insert_text(self):
+        from cloakbrowser.human.keyboard import human_type
+        from cloakbrowser.human.config import resolve_config
+        from unittest.mock import MagicMock
+
+        cfg = resolve_config("default", {"mistype_chance": 0})
+        page = MagicMock()
+        raw = MagicMock()
+
+        inserted = []
+        raw.down = MagicMock()
+        raw.up = MagicMock()
+        raw.insert_text = MagicMock(side_effect=lambda t: inserted.append(t))
+
+        human_type(page, raw, "你好", cfg)
+
+        assert "".join(inserted) == "你好"
+
+    def test_mistype_only_ascii(self):
+        from cloakbrowser.human.keyboard import human_type
+        from cloakbrowser.human.config import resolve_config
+        from unittest.mock import MagicMock
+
+        cfg = resolve_config("default", {"mistype_chance": 1.0})
+        page = MagicMock()
+        raw = MagicMock()
+
+        down_keys = []
+        raw.down = MagicMock(side_effect=lambda k: down_keys.append(k))
+        raw.up = MagicMock()
+        raw.insert_text = MagicMock()
+
+        human_type(page, raw, "AБ", cfg)
+
+        assert "Backspace" in down_keys
+
+    def test_no_error_on_cyrillic(self):
+        from cloakbrowser.human.keyboard import human_type
+        from cloakbrowser.human.config import resolve_config
+        from unittest.mock import MagicMock
+
+        cfg = resolve_config("default", {"mistype_chance": 0})
+        page = MagicMock()
+        raw = MagicMock()
+        raw.down = MagicMock()
+        raw.up = MagicMock()
+        raw.insert_text = MagicMock()
+
+        # Should not raise
+        human_type(page, raw, "Тест кириллицы", cfg)
+
+
+class TestNonAsciiKeyboardAsync:
+    @pytest.mark.asyncio
+    async def test_async_cyrillic_uses_insert_text(self):
+        from cloakbrowser.human.keyboard_async import async_human_type
+        from cloakbrowser.human.config import resolve_config
+        from unittest.mock import MagicMock, AsyncMock
+
+        cfg = resolve_config("default", {"mistype_chance": 0})
+        page = MagicMock()
+        raw = MagicMock()
+
+        inserted = []
+        raw.down = AsyncMock()
+        raw.up = AsyncMock()
+        raw.insert_text = AsyncMock(side_effect=lambda t: inserted.append(t))
+
+        await async_human_type(page, raw, "Привет", cfg)
+
+        assert "".join(inserted) == "Привет"
+
+
+
 # =========================================================================
 # SLOW TESTS — require browser (skipped in CI unless pytest -m slow)
 # =========================================================================