feat(core/python): create testing module in trezorlib

diff --git a/python/.changelog.d/6994.added b/python/.changelog.d/6994.added
new file mode 100644
index 00000000..da4adf45
--- /dev/null
+++ b/python/.changelog.d/6994.added
@@ -0,0 +1 @@
+Add testing module to trezorlib.
diff --git a/python/src/trezorlib/testing/__init__.py b/python/src/trezorlib/testing/__init__.py
new file mode 100644
index 00000000..e69de29b
diff --git a/python/src/trezorlib/testing/common.py b/python/src/trezorlib/testing/common.py
new file mode 100644
index 00000000..5c883e32
--- /dev/null
+++ b/python/src/trezorlib/testing/common.py
@@ -0,0 +1,93 @@
+from typing import TYPE_CHECKING, Generator, Union
+
+from .. import messages
+from ..debuglink import LayoutType
+
+if TYPE_CHECKING:
+
+    from ..debuglink import DebugLink
+
+
+BRGeneratorType = Generator[None, messages.ButtonRequest, None]
+
+PRIVATE_KEYS_DEV = [byte * 32 for byte in (b"\xdd", b"\xde", b"\xdf")]
+
+
+def compact_size(n: int) -> bytes:
+    """
+    Encode an integer using Bitcoin's compact size format.
+
+    Args:
+        n (int): The integer to encode.
+
+    Returns:
+        bytes: The encoded integer.
+
+    Raises:
+        ValueError: If n is not in the range 0..2^64-1.
+    """
+    if n < 0 or n > 0xFFFF_FFFF_FFFF_FFFF:
+        raise ValueError("compact_size supports integers in range 0..2^64-1")
+    if n < 253:
+        return n.to_bytes(1, "little")
+    elif n < 0x1_0000:
+        return bytes([253]) + n.to_bytes(2, "little")
+    elif n < 0x1_0000_0000:
+        return bytes([254]) + n.to_bytes(4, "little")
+    else:
+        return bytes([255]) + n.to_bytes(8, "little")
+
+
+def get_text_possible_pagination(debug: "DebugLink", br: messages.ButtonRequest) -> str:
+    """
+    Read all text content from the device, handling possible pagination.
+
+    Args:
+        debug (DebugLink): The debug link to interact with the device.
+        br (ButtonRequest): The button request containing pagination info.
+
+    Returns:
+        str: The concatenated text from all pages.
+    """
+    text = debug.read_layout().text_content()
+    if br.pages is not None:
+        for _ in range(br.pages - 1):
+            if debug.layout_type is LayoutType.Eckhart:
+                debug.click(debug.screen_buttons.ok())
+            else:
+                debug.swipe_up()
+            text += " "
+            text += debug.read_layout().text_content()
+    return text
+
+
+def swipe_if_necessary(
+    debug: "DebugLink", br_code: Union[messages.ButtonRequestType, None] = None
+) -> BRGeneratorType:
+    """
+    Generator that swipes through pages if necessary, based on button request code.
+
+    Args:
+        debug (DebugLink): The debug link to interact with the device.
+        br_code (ButtonRequestType or None): Expected button request code.
+
+    Yields:
+        ButtonRequest: The button request to process.
+    """
+    br = yield
+    if br_code is not None:
+        assert br.code == br_code
+    swipe_till_the_end(debug, br)
+
+
+def swipe_till_the_end(debug: "DebugLink", br: messages.ButtonRequest) -> None:
+    """
+    Swipe through all pages until the end.
+
+    Args:
+        debug (DebugLink): The debug link to interact with the device.
+        br (ButtonRequest): The button request containing pagination info.
+    """
+    if br.pages is not None:
+        for _ in range(br.pages - 1):
+            debug.swipe_up()
diff --git a/python/src/trezorlib/testing/device_handler.py b/python/src/trezorlib/testing/device_handler.py
new file mode 100644
index 00000000..cb3970f0
--- /dev/null
+++ b/python/src/trezorlib/testing/device_handler.py
@@ -0,0 +1,269 @@
+from __future__ import annotations
+
+import typing as t
+from concurrent.futures import ThreadPoolExecutor
+
+import typing_extensions as tx
+
+from ..messages import DebugWaitType
+from ..transport import udp
+
+if t.TYPE_CHECKING:
+    from .._internal.emulator import Emulator
+    from ..client import Session
+    from ..debuglink import DebugLink
+    from ..debuglink import TrezorTestContext as Client
+    from ..messages import Features
+
+    P = tx.ParamSpec("P")
+
+
+udp.SOCKET_TIMEOUT = 0.1
+
+
+class NullUI:
+    """
+    A mock UI class for testing purposes.
+
+    Provides static methods that do nothing or raise errors, simulating
+    a UI interface for use in tests where UI interaction is not expected.
+    """
+
+    @staticmethod
+    def clear(*args: t.Any, **kwargs: t.Any) -> None:
+        """
+        Mock clear method. Does nothing.
+        """
+
+    @staticmethod
+    def button_request(code: t.Any) -> None:
+        """
+        Mock button request handler. Does nothing.
+        """
+
+    @staticmethod
+    def get_pin(code: t.Any = None) -> None:
+        """
+        Mock PIN entry handler.
+
+        Raises:
+            NotImplementedError: Always, as PIN entry should not be used with NullUI.
+        """
+        raise NotImplementedError("NullUI should not be used with T1")
+
+
+class BackgroundDeviceHandler:
+    """
+    Handles background operations and UI simulation for device testing.
+
+    Manages asynchronous device interactions, session handling, and UI events
+    in a testing environment. Ensures that device tasks are properly started,
+    awaited, and finalized.
+    """
+
+    _pool = ThreadPoolExecutor()
+
+    def __init__(self, client: "Client", nowait: bool = False) -> None:
+        """
+        Initialize the handler with a client.
+
+        Args:
+            client: The client object to interact with.
+            nowait: If True, do not wait for tasks to finish before returning.
+        """
+        self._configure_client(client)
+        self.task = None
+        self.nowait = nowait
+
+    def _configure_client(self, client: "Client") -> None:
+        """
+        Configure the client for background handling and set up the mock UI.
+
+        Args:
+            client: The client object to configure.
+        """
+        self.client = client
+        self.client.ui = NullUI()  # pyright: ignore [reportAttributeAccessIssue]
+        self.client.app.button_callback = self.client.ui.button_request
+        self.client.debug.input_wait_type = DebugWaitType.CURRENT_LAYOUT
+
+    def get_session(self, *args: t.Any, **kwargs: t.Any) -> None:
+        """
+        Start a background task to get a session, waiting for a layout change.
+
+        Raises:
+            RuntimeError: If a previous task is still running.
+        """
+        if self.task is not None:
+            raise RuntimeError("Wait for previous task first")
+
+        with self.debuglink().wait_for_layout_change():
+            self.task = self._pool.submit(self.client.get_session, *args, **kwargs)
+
+    def run_with_session(
+        self,
+        function: t.Callable[tx.Concatenate["Session", P], t.Any],
+        seedless: bool = False,
+        *args: P.args,
+        **kwargs: P.kwargs,
+    ) -> None:
+        """
+        Run a function that interacts with a device session in the background.
+
+        Ensures the UI is updated before returning.
+
+        Args:
+            function: The function to run with the session.
+            seedless: If True, use a seedless session.
+            *args: Additional positional arguments for the function.
+            **kwargs: Additional keyword arguments for the function.
+
+        Raises:
+            RuntimeError: If a previous task is still running.
+        """
+        if self.task is not None:
+            raise RuntimeError("Wait for previous task first")
+
+        def task_function(*args: t.Any, **kwargs: t.Any) -> t.Any:
+            if seedless:
+                session = self.client.get_seedless_session()
+            else:
+                session = self.client.get_session()
+            return function(session, *args, **kwargs)
+
+        # wait for the first UI change triggered by the task running in the background
+        with self.debuglink().wait_for_layout_change():
+            self.task = self._pool.submit(task_function, *args, **kwargs)
+
+    def run_with_provided_session(
+        self,
+        session: "Session",
+        function: t.Callable[tx.Concatenate["Session", P], t.Any],
+        *args: P.args,
+        **kwargs: P.kwargs,
+    ) -> None:
+        """
+        Run a function with a provided session in the background.
+
+        Ensures the UI is updated before returning.
+
+        Args:
+            session: The session to use.
+            function: The function to run.
+            *args: Additional positional arguments for the function.
+            **kwargs: Additional keyword arguments for the function.
+
+        Raises:
+            RuntimeError: If a previous task is still running.
+        """
+        if self.task is not None:
+            raise RuntimeError("Wait for previous task first")
+
+        # wait for the first UI change triggered by the task running in the background
+        with self.debuglink().wait_for_layout_change():
+            self.task = self._pool.submit(function, session, *args, **kwargs)
+
+    def kill_task(self) -> None:
+        """
+        Kill the currently running background task, if any.
+
+        Closes the client transport and waits for the task to finish.
+        """
+        if self.task is not None:
+            # Force close the transport, which should raise an exception in a client
+            # waiting on IO. Does not work over Bridge, because bridge doesn't have
+            # a close() method.
+            self.client.transport.close()
+            try:
+                self.task.result(timeout=1)
+            except Exception:
+                pass
+        self.task = None
+
+    def restart(self, emulator: "Emulator") -> None:
+        """
+        Restart the emulator and reconfigure the client.
+
+        Args:
+            emulator: The emulator to restart.
+        """
+        # TODO handle actual restart as well
+        self.kill_task()
+        emulator.restart()
+        self._configure_client(emulator.client)
+
+    def result(self, timeout: float | None = None) -> t.Any:
+        """
+        Get the result of the background task.
+
+        Args:
+            timeout: Optional timeout in seconds.
+
+        Returns:
+            The result of the task.
+
+        Raises:
+            RuntimeError: If no task is running.
+        """
+        if self.task is None:
+            raise RuntimeError("No task running")
+        try:
+            return self.task.result(timeout=timeout)
+        finally:
+            self.task = None
+
+    def features(self) -> "Features":
+        """
+        Refresh and return the client's features.
+
+        Returns:
+            Features: The client's features.
+
+        Raises:
+            RuntimeError: If a task is running.
+        """
+        if self.task is not None:
+            raise RuntimeError("Cannot query features while task is running")
+        self.client.refresh_features()
+        return self.client.features
+
+    def debuglink(self) -> "DebugLink":
+        """
+        Get the debuglink for the client.
+
+        Returns:
+            DebugLink: The client's debuglink.
+        """
+        return self.client.debug
+
+    def check_finalize(self) -> bool:
+        """
+        Ensure all tasks are finalized.
+
+        Returns:
+            bool: True if no task was running, False if a task was killed.
+        """
+        if self.task is not None:
+            self.kill_task()
+            return False
+        return True
+
+    def __enter__(self) -> "BackgroundDeviceHandler":
+        """
+        Enter the context manager.
+
+        Returns:
+            BackgroundDeviceHandler: self
+        """
+        return self
+
+    def __exit__(self, exc_type: t.Any, exc_value: t.Any, traceback: t.Any) -> None:
+        """
+        Exit the context manager, ensuring all tasks are finalized.
+
+        Raises:
+            RuntimeError: If exiting while a task is unfinished.
+        """
+        finalized_ok = self.check_finalize()
+        if exc_type is None and not finalized_ok:
+            raise RuntimeError("Exit while task is unfinished")
diff --git a/python/src/trezorlib/testing/translations.py b/python/src/trezorlib/testing/translations.py
new file mode 100644
index 00000000..9234ec15
--- /dev/null
+++ b/python/src/trezorlib/testing/translations.py
@@ -0,0 +1,369 @@
+from __future__ import annotations
+
+import json
+import os
+import re
+import threading
+import typing as t
+import warnings
+from hashlib import sha256
+from pathlib import Path
+
+from .. import cosi, device, models
+from .._internal import translations
+from ..debuglink import DebugSession, LayoutType
+from . import common
+
+HERE = Path(__file__).resolve().parent
+ROOT = HERE.parent.parent.parent.parent
+
+
+def get_translations_dir() -> Path:
+    env_dir = os.environ.get("TREZOR_TRANSLATIONS_DIR")
+    if env_dir:
+        env_dir = Path(env_dir)
+        if env_dir.is_dir():
+            return env_dir
+        raise FileNotFoundError(
+            f"TREZOR_TRANSLATIONS_DIR is set to '{env_dir}' but it is not a valid directory."
+        )
+    else:
+        # Legacy: translations living in the firmware repo
+        legacy = ROOT / "core" / "translations"
+        if legacy.is_dir():
+            return legacy
+
+        raise FileNotFoundError(
+            "Translations directory not found. Set TREZOR_TRANSLATIONS_DIR environment variable."
+        )
+
+
+TRANSLATIONS_DIR = get_translations_dir()
+FONTS_DIR = TRANSLATIONS_DIR / "fonts"
+ORDER_FILE = TRANSLATIONS_DIR / "order.json"
+
+LANGUAGES = [file.stem for file in TRANSLATIONS_DIR.glob("??.json")]
+
+_CURRENT_TRANSLATION = threading.local()
+
+
+def prepare_blob(
+    lang_or_def: translations.JsonDef | Path | str,
+    model: models.TrezorModel,
+    version: translations.VersionTuple | tuple[int, int, int] | None = None,
+) -> translations.TranslationsBlob:
+    """
+    Prepare a translation blob for a given language and device model.
+
+    Args:
+        lang_or_def: Language identifier, JSON definition, or path to JSON file.
+        model: Trezor device model.
+        version: Optional version tuple.
+
+    Returns:
+        TranslationsBlob: The prepared translation blob.
+    """
+    order = translations.order_from_json(json.loads(ORDER_FILE.read_text()))
+    if isinstance(lang_or_def, str):
+        lang_or_def = get_lang_json(lang_or_def)
+    if isinstance(lang_or_def, Path):
+        lang_or_def = t.cast(translations.JsonDef, json.loads(lang_or_def.read_text()))
+
+    # generate raw blob
+    if version is None:
+        version = translations.version_from_json(lang_or_def["header"]["version"])
+    elif len(version) == 3:
+        # version coming from client object does not have build item
+        version = *version, 0
+    return translations.blob_from_defs(lang_or_def, order, model, version, FONTS_DIR)
+
+
+def sign_blob(blob: translations.TranslationsBlob) -> bytes:
+    """
+    Sign a translation blob using the developer private keys.
+
+    Args:
+        blob: The translation blob to sign.
+
+    Returns:
+        bytes: The signed blob as bytes.
+    """
+    # build 0-item Merkle proof
+    digest = sha256(b"\x00" + blob.header_bytes).digest()
+    signature = cosi.sign_with_privkeys(digest, common.PRIVATE_KEYS_DEV)
+    blob.proof = translations.Proof(
+        merkle_proof=[],
+        sigmask=0b111,
+        signature=signature,
+    )
+    return blob.build()
+
+
+def build_and_sign_blob(
+    lang_or_def: translations.JsonDef | Path | str,
+    session: DebugSession,
+) -> bytes:
+    """
+    Prepare and sign a translation blob for a given language and session.
+
+    Args:
+        lang_or_def: Language identifier, JSON definition, or path to JSON file.
+        session: DebugSession object.
+
+    Returns:
+        bytes: The signed translation blob.
+    """
+    blob = prepare_blob(lang_or_def, session.model, session.version)
+    return sign_blob(blob)
+
+
+def set_language(session: DebugSession, lang: str, *, force: bool = False) -> None:
+    """
+    Set the device language for a given session.
+
+    Args:
+        session: DebugSession object.
+        lang: Language code (e.g., 'en', 'cs').
+        force: If True, force setting the language even if already set.
+    """
+    if lang.startswith("en"):
+        language_data = b""
+    else:
+        language_data = build_and_sign_blob(lang, session)
+    with session.test_ctx:
+        language = session.features.language
+        if language is None or not language.startswith(lang) or force:
+            device.change_language(session, language_data)
+    _CURRENT_TRANSLATION.LAYOUT = session.layout_type
+    _CURRENT_TRANSLATION.TR = TRANSLATIONS[lang]
+
+
+def check_language(session: DebugSession, lang: str) -> None:
+    """
+    Assert that the device language matches the expected language.
+
+    Args:
+        session: DebugSession object.
+        lang: Expected language code.
+
+    Raises:
+        RuntimeError: If the device language does not match.
+    """
+    with session.test_ctx:
+        language = session.features.language
+        assert isinstance(language, str)
+        if not language.startswith(lang):
+            raise RuntimeError(
+                f"Incompatible language on device: expected '{lang}', got '{language}'"
+            )
+    _CURRENT_TRANSLATION.LAYOUT = session.layout_type
+    _CURRENT_TRANSLATION.TR = TRANSLATIONS[lang]
+
+
+def get_language() -> str:
+    """
+    Get the current language code.
+
+    Returns:
+        str: The current language code.
+    """
+    for lang in LANGUAGES:
+        if _CURRENT_TRANSLATION.TR == TRANSLATIONS[lang]:
+            return lang
+    return "en"
+
+
+def get_lang_json(lang: str) -> translations.JsonDef:
+    """
+    Load the JSON definition for a given language.
+
+    Args:
+        lang: Language code.
+
+    Returns:
+        JsonDef: The loaded JSON definition.
+
+    Raises:
+        AssertionError: If the language is not available.
+    """
+    assert lang in LANGUAGES
+    lang_json = json.loads((TRANSLATIONS_DIR / f"{lang}.json").read_text())
+    if (fonts_safe3 := lang_json.get("fonts", {}).get("##Safe3")) is not None:
+        lang_json["fonts"]["T2B1"] = fonts_safe3
+        lang_json["fonts"]["T3B1"] = fonts_safe3
+    return lang_json
+
+
+class Translation:
+    """
+    Represents a translation for a specific language.
+
+    Provides methods for translating keys, formatting, and generating regex patterns.
+    """
+
+    FORMAT_STR_RE = re.compile(r"\\{\d+\\}")
+
+    def __init__(self, lang: str) -> None:
+        """
+        Initialize a Translation object.
+
+        Args:
+            lang: Language code.
+        """
+        self.lang = lang
+        self.lang_json = get_lang_json(lang)
+
+    @property
+    def translations(self) -> dict[str, str | dict[str, str]]:
+        """
+        Get the translations dictionary.
+
+        Returns:
+            dict[str, str | dict[str, str]]: The translations dictionary.
+        """
+        return self.lang_json["translations"]
+
+    def _translate_raw(self, key: str, _stacklevel: int = 0) -> str:
+        tr = self.translations.get(key)
+        # Emulate firmware behaviour: English strings are used instead of missing & empty strings
+        if tr:
+            # Handle layout-specific translations
+            if isinstance(tr, dict) and hasattr(_CURRENT_TRANSLATION, "LAYOUT"):
+                # Try to get translation for current layout
+                layout_name = _CURRENT_TRANSLATION.LAYOUT.name
+                if layout_name in tr:
+                    return tr[layout_name]
+                # Fall back to any available translation if no match for current layout
+                return next(iter(tr.values()))
+            elif isinstance(tr, str):
+                return tr
+            else:
+                raise ValueError(f"Invalid translation value for key '{key}'")
+        if self.lang != "en":
+            # check if the key exists in English first
+            retval = TRANSLATIONS["en"]._translate_raw(key)
+            # if not, a KeyError was raised so we fall through.
+            # otherwise, warn that the key is untranslated in target language.
+            warnings.warn(
+                f"Translation key '{key}' not found in '{self.lang}' translation file",
+                stacklevel=_stacklevel + 2,
+            )
+            return retval
+        raise KeyError(key)
+
+    def translate(self, key: str, _stacklevel: int = 0) -> str:
+        """
+        Get the translated string for a key.
+
+        Args:
+            key: Translation key.
+            _stacklevel: Internal stacklevel for warnings.
+
+        Returns:
+            str: The translated string.
+        """
+        return self._translate_raw(key, _stacklevel=_stacklevel + 1).strip()
+
+    def as_regexp(self, key: str, _stacklevel: int = 0) -> re.Pattern:
+        """
+        Get a regular expression pattern for a translation key.
+
+        Args:
+            key: Translation key.
+            _stacklevel: Internal stacklevel for warnings.
+
+        Returns:
+            re.Pattern: The compiled regular expression.
+        """
+        tr = self.translate(key, _stacklevel=_stacklevel + 1)
+        re_safe = re.escape(tr)
+        return re.compile(self.FORMAT_STR_RE.sub(r".*?", re_safe))
+
+    def format(self, key: str, *args: t.Any, **kwargs: t.Any) -> str:
+        """
+        Format a translation string with arguments.
+
+        Args:
+            key: Translation key.
+            *args: Positional arguments for formatting.
+            **kwargs: Keyword arguments for formatting.
+
+        Returns:
+            str: The formatted translation string.
+
+        Raises:
+            ValueError: If formatting fails.
+        """
+        tr = self.translate(key)
+        try:
+            return tr.format(*args, **kwargs)
+        except (KeyError, IndexError) as e:
+            raise ValueError(
+                f"Failed to format translation '{key}' with args={args}, kwargs={kwargs}: {e}"
+            ) from e
+
+
+TRANSLATIONS = {lang: Translation(lang) for lang in LANGUAGES}
+_CURRENT_TRANSLATION.TR = TRANSLATIONS["en"]
+_CURRENT_TRANSLATION.LAYOUT = LayoutType.Bolt
+
+
+def translate(key: str, _stacklevel: int = 0) -> str:
+    """
+    Translate a key using the current translation.
+
+    Args:
+        key: Translation key.
+        _stacklevel: Internal stacklevel for warnings.
+
+    Returns:
+        str: The translated string.
+    """
+    return _CURRENT_TRANSLATION.TR.translate(key, _stacklevel=_stacklevel + 1)
+
+
+def regexp(key: str) -> re.Pattern:
+    """
+    Get a regular expression pattern for a translation key.
+
+    Args:
+        key: Translation key.
+
+    Returns:
+        re.Pattern: The compiled regular expression.
+    """
+    return _CURRENT_TRANSLATION.TR.as_regexp(key, _stacklevel=1)
+
+
+def format(key: str, *args: t.Any, **kwargs: t.Any) -> str:
+    """
+    Format a translation string with arguments.
+
+    Args:
+        key: Translation key.
+        *args: Positional arguments for formatting.
+        **kwargs: Keyword arguments for formatting.
+
+    Returns:
+        str: The formatted translation string.
+    """
+    return _CURRENT_TRANSLATION.TR.format(key, *args, **kwargs)
+
+
+def __getattr__(key: str) -> str:
+    """
+    Allow attribute-style access to translations.
+
+    Args:
+        key: Translation key.
+
+    Returns:
+        str: The translated string.
+
+    Raises:
+        AttributeError: If the translation key is not found.
+    """
+    try:
+        return translate(key, _stacklevel=1)
+    except KeyError as e:
+        raise AttributeError(f"Translation key '{key}' not found") from e
diff --git a/tests/click_tests/common.py b/tests/click_tests/common.py
index e9c0e4a9..492f5850 100644
--- a/tests/click_tests/common.py
+++ b/tests/click_tests/common.py
@@ -5,8 +5,7 @@ import typing as t
 from enum import Enum
 
 from trezorlib.debuglink import LayoutType
-
-from .. import translations as TR
+from trezorlib.testing import translations as TR
 
 if t.TYPE_CHECKING:
     from trezorlib.debuglink import DebugLink, LayoutContent
diff --git a/tests/click_tests/device_menu/common.py b/tests/click_tests/device_menu/common.py
index d70a7e39..4b2cfa7c 100644
--- a/tests/click_tests/device_menu/common.py
+++ b/tests/click_tests/device_menu/common.py
@@ -18,14 +18,12 @@ from enum import Enum, auto
 from typing import TYPE_CHECKING, Callable
 
 from trezorlib.messages import BackupAvailability
-
-from ... import translations as TR
+from trezorlib.testing import translations as TR
 
 if TYPE_CHECKING:
     from trezorlib.debuglink import DebugLink
     from trezorlib.messages import Features
-
-    from ...device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 PIN4 = "1234"
 
diff --git a/tests/click_tests/device_menu/test_auto_lock.py b/tests/click_tests/device_menu/test_auto_lock.py
index ef432a72..a61171b5 100644
--- a/tests/click_tests/device_menu/test_auto_lock.py
+++ b/tests/click_tests/device_menu/test_auto_lock.py
@@ -18,7 +18,8 @@ from typing import TYPE_CHECKING
 
 import pytest
 
-from ... import translations as TR
+from trezorlib.testing import translations as TR
+
 from ..test_pin import PIN4
 from .common import (
     Menu,
@@ -33,8 +34,7 @@ from .common import (
 if TYPE_CHECKING:
     from trezorlib.debuglink import DebugLink
     from trezorlib.messages import Features
-
-    from ...device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 # Trezor Safe 7 only
 pytestmark = [pytest.mark.models("eckhart")]
 
diff --git a/tests/click_tests/device_menu/test_check_backup.py b/tests/click_tests/device_menu/test_check_backup.py
index 20f8d524..efbe4213 100644
--- a/tests/click_tests/device_menu/test_check_backup.py
+++ b/tests/click_tests/device_menu/test_check_backup.py
@@ -19,8 +19,8 @@ from typing import TYPE_CHECKING
 import pytest
 
 from trezorlib import messages
+from trezorlib.testing import translations as TR
 
-from ... import translations as TR
 from ...common import MNEMONIC12
 from .common import (
     Menu,
@@ -34,8 +34,7 @@ from .common import (
 if TYPE_CHECKING:
     from trezorlib.debuglink import DebugLink
     from trezorlib.messages import Features
-
-    from ...device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 # Trezor Safe 7 only
 pytestmark = [pytest.mark.models("eckhart")]
 
diff --git a/tests/click_tests/device_menu/test_device_settings.py b/tests/click_tests/device_menu/test_device_settings.py
index 8f683e7e..017b300f 100644
--- a/tests/click_tests/device_menu/test_device_settings.py
+++ b/tests/click_tests/device_menu/test_device_settings.py
@@ -18,14 +18,14 @@ from typing import TYPE_CHECKING
 
 import pytest
 
-from ... import translations as TR
+from trezorlib.testing import translations as TR
+
 from .common import Menu, assert_device_screen, close_device_menu, open_device_menu
 
 if TYPE_CHECKING:
     from trezorlib.debuglink import DebugLink
     from trezorlib.messages import Features
-
-    from ...device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 # Trezor Safe 7 only
 pytestmark = [pytest.mark.models("eckhart")]
 
diff --git a/tests/click_tests/device_menu/test_label.py b/tests/click_tests/device_menu/test_label.py
index 1a70d012..91120bb8 100644
--- a/tests/click_tests/device_menu/test_label.py
+++ b/tests/click_tests/device_menu/test_label.py
@@ -18,7 +18,8 @@ from typing import TYPE_CHECKING
 
 import pytest
 
-from ... import translations as TR
+from trezorlib.testing import translations as TR
+
 from ..common import KeyboardCategory, delete_char, go_to_category, press_char
 from .common import (
     Menu,
@@ -31,8 +32,7 @@ from .common import (
 if TYPE_CHECKING:
     from trezorlib.debuglink import DebugLink
     from trezorlib.messages import Features
-
-    from ...device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 # Trezor Safe 7 only
 pytestmark = [pytest.mark.models("eckhart")]
 
diff --git a/tests/click_tests/device_menu/test_notifications.py b/tests/click_tests/device_menu/test_notifications.py
index addfb235..3c7602f4 100644
--- a/tests/click_tests/device_menu/test_notifications.py
+++ b/tests/click_tests/device_menu/test_notifications.py
@@ -19,8 +19,8 @@ from typing import TYPE_CHECKING
 import pytest
 
 from trezorlib import messages
+from trezorlib.testing import translations as TR
 
-from ... import translations as TR
 from ..test_pin import PIN4, _assert_pin_entry, _enter_two_times
 from .common import (
     Menu,
@@ -33,7 +33,7 @@ from .common import (
 
 if TYPE_CHECKING:
 
-    from ...device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 # Trezor Safe 7 only
 pytestmark = [pytest.mark.models("eckhart")]
 
diff --git a/tests/click_tests/device_menu/test_pin.py b/tests/click_tests/device_menu/test_pin.py
index 1a538aa9..af2163d4 100644
--- a/tests/click_tests/device_menu/test_pin.py
+++ b/tests/click_tests/device_menu/test_pin.py
@@ -19,7 +19,8 @@ from typing import TYPE_CHECKING
 
 import pytest
 
-from ... import translations as TR
+from trezorlib.testing import translations as TR
+
 from ..common import go_next
 from ..test_pin import (
     PIN1,
@@ -43,8 +44,7 @@ from .common import (
 if TYPE_CHECKING:
     from trezorlib.debuglink import DebugLink
     from trezorlib.messages import Features
-
-    from ...device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 # Trezor Safe 7 only
 pytestmark = [pytest.mark.models("eckhart")]
 
diff --git a/tests/click_tests/device_menu/test_traverse_menu.py b/tests/click_tests/device_menu/test_traverse_menu.py
index dc40571b..4f0b615e 100644
--- a/tests/click_tests/device_menu/test_traverse_menu.py
+++ b/tests/click_tests/device_menu/test_traverse_menu.py
@@ -21,7 +21,7 @@ import pytest
 from .common import PIN4, Menu, close_device_menu, enter_pin, open_device_menu
 
 if TYPE_CHECKING:
-    from ...device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 # Trezor Safe 7 only
 pytestmark = [pytest.mark.models("eckhart")]
diff --git a/tests/click_tests/device_menu/test_wipe_code.py b/tests/click_tests/device_menu/test_wipe_code.py
index 0b68e319..33e21032 100644
--- a/tests/click_tests/device_menu/test_wipe_code.py
+++ b/tests/click_tests/device_menu/test_wipe_code.py
@@ -18,7 +18,8 @@ from typing import TYPE_CHECKING
 
 import pytest
 
-from ... import translations as TR
+from trezorlib.testing import translations as TR
+
 from ..common import go_next
 from ..test_pin import (
     PIN1,
@@ -42,8 +43,7 @@ from .test_pin import Situation, prepare_pin_dialogue
 
 if TYPE_CHECKING:
     from trezorlib.debuglink import DebugLink
-
-    from ...device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 # Trezor Safe 7 only
 pytestmark = [pytest.mark.models("eckhart")]
 
diff --git a/tests/click_tests/device_menu/test_wipe_device.py b/tests/click_tests/device_menu/test_wipe_device.py
index 7f7ffbca..4bf87458 100644
--- a/tests/click_tests/device_menu/test_wipe_device.py
+++ b/tests/click_tests/device_menu/test_wipe_device.py
@@ -18,11 +18,12 @@ from typing import TYPE_CHECKING
 
 import pytest
 
-from ... import translations as TR
+from trezorlib.testing import translations as TR
+
 from .common import PIN4, Menu, enter_pin, open_device_menu
 
 if TYPE_CHECKING:
-    from ...device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 # Trezor Safe 7 only
 pytestmark = [pytest.mark.models("eckhart")]
diff --git a/tests/click_tests/recovery.py b/tests/click_tests/recovery.py
index 7504524d..4f4f6cbe 100644
--- a/tests/click_tests/recovery.py
+++ b/tests/click_tests/recovery.py
@@ -1,8 +1,8 @@
 from typing import TYPE_CHECKING
 
 from trezorlib.debuglink import LayoutType
+from trezorlib.testing import translations as TR
 
-from .. import translations as TR
 from ..recovery_helpers import navigate_to_keyboard
 from .common import go_next
 
diff --git a/tests/click_tests/reset.py b/tests/click_tests/reset.py
index 4f3b139e..86d9ae62 100644
--- a/tests/click_tests/reset.py
+++ b/tests/click_tests/reset.py
@@ -4,8 +4,7 @@ from typing import TYPE_CHECKING
 from shamir_mnemonic import shamir  # type: ignore
 
 from trezorlib.debuglink import LayoutType
-
-from .. import translations as TR
+from trezorlib.testing import translations as TR
 
 if TYPE_CHECKING:
     from trezorlib.debuglink import DebugLink
diff --git a/tests/click_tests/test_autolock.py b/tests/click_tests/test_autolock.py
index 689459c2..cc79b279 100644
--- a/tests/click_tests/test_autolock.py
+++ b/tests/click_tests/test_autolock.py
@@ -24,10 +24,10 @@ from trezorlib import btc, device, exceptions, messages
 from trezorlib.client import PassphraseSetting
 from trezorlib.debuglink import DebugLink, LayoutType
 from trezorlib.protobuf import MessageType
+from trezorlib.testing import translations as TR
 from trezorlib.tools import parse_path
 
 from .. import common
-from .. import translations as TR
 from ..device_tests.bitcoin.coinjoin_req import make_coinjoin_request
 from ..tx_cache import TxCache
 from . import recovery
@@ -35,8 +35,7 @@ from .common import go_next, unlock_gesture
 
 if TYPE_CHECKING:
     from trezorlib.debuglink import LayoutContent
-
-    from ..device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 TX_CACHE_MAINNET = TxCache("Bitcoin")
 TX_CACHE_TESTNET = TxCache("Testnet")
diff --git a/tests/click_tests/test_backup_slip39_custom.py b/tests/click_tests/test_backup_slip39_custom.py
index e380c29f..e874485c 100644
--- a/tests/click_tests/test_backup_slip39_custom.py
+++ b/tests/click_tests/test_backup_slip39_custom.py
@@ -20,13 +20,13 @@ import pytest
 
 from trezorlib import device, messages
 from trezorlib.debuglink import LayoutType
+from trezorlib.testing import translations as TR
 
-from .. import translations as TR
 from ..common import EXTERNAL_ENTROPY, MOCK_GET_ENTROPY, generate_entropy
 from . import reset
 
 if TYPE_CHECKING:
-    from ..device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 
 pytestmark = pytest.mark.models("core")
diff --git a/tests/click_tests/test_lock.py b/tests/click_tests/test_lock.py
index 38fb91e1..bacd4bca 100644
--- a/tests/click_tests/test_lock.py
+++ b/tests/click_tests/test_lock.py
@@ -23,7 +23,7 @@ from trezorlib import messages, models
 from trezorlib.debuglink import LayoutType
 
 if TYPE_CHECKING:
-    from ..device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 
 PIN4 = "1234"
diff --git a/tests/click_tests/test_passphrase_bde.py b/tests/click_tests/test_passphrase_bde.py
index 58e20fd4..50327a62 100644
--- a/tests/click_tests/test_passphrase_bde.py
+++ b/tests/click_tests/test_passphrase_bde.py
@@ -36,8 +36,7 @@ from .common import (  # KEYBOARD_CATEGORY,
 
 if TYPE_CHECKING:
     from trezorlib.debuglink import DebugLink
-
-    from ..device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 
 pytestmark = pytest.mark.models("t2t1", "delizia", "eckhart")
diff --git a/tests/click_tests/test_pin.py b/tests/click_tests/test_pin.py
index e13d4f4b..89085a07 100644
--- a/tests/click_tests/test_pin.py
+++ b/tests/click_tests/test_pin.py
@@ -23,14 +23,13 @@ import pytest
 
 from trezorlib import device, exceptions
 from trezorlib.debuglink import DisplayStyle, LayoutType
+from trezorlib.testing import translations as TR
 
-from .. import translations as TR
 from .common import go_next, navigate_to_action_and_press
 
 if TYPE_CHECKING:
     from trezorlib.debuglink import DebugLink
-
-    from ..device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 
 pytestmark = pytest.mark.models("core")
diff --git a/tests/click_tests/test_recovery.py b/tests/click_tests/test_recovery.py
index be3136f9..f83f71e0 100644
--- a/tests/click_tests/test_recovery.py
+++ b/tests/click_tests/test_recovery.py
@@ -29,8 +29,7 @@ from .test_autolock import PIN4, set_autolock_delay, unlock_dry_run
 
 if TYPE_CHECKING:
     from trezorlib.debuglink import DebugLink
-
-    from ..device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 
 pytestmark = pytest.mark.models("core")
diff --git a/tests/click_tests/test_repeated_backup.py b/tests/click_tests/test_repeated_backup.py
index f7a0ef23..e71fa169 100644
--- a/tests/click_tests/test_repeated_backup.py
+++ b/tests/click_tests/test_repeated_backup.py
@@ -19,14 +19,14 @@ from typing import TYPE_CHECKING
 import pytest
 
 from trezorlib import device, exceptions, messages
+from trezorlib.testing import translations as TR
 
-from .. import translations as TR
 from ..common import MOCK_GET_ENTROPY, LayoutType
 from . import recovery, reset
 from .common import go_next
 
 if TYPE_CHECKING:
-    from ..device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 
 pytestmark = pytest.mark.models("core")
diff --git a/tests/click_tests/test_reset_bip39.py b/tests/click_tests/test_reset_bip39.py
index be257aaf..cfce2503 100644
--- a/tests/click_tests/test_reset_bip39.py
+++ b/tests/click_tests/test_reset_bip39.py
@@ -19,14 +19,14 @@ from typing import TYPE_CHECKING
 import pytest
 
 from trezorlib import device, messages
+from trezorlib.testing import translations as TR
 
-from .. import translations as TR
 from ..common import MOCK_GET_ENTROPY
 from . import reset
 from .common import LayoutType, go_next
 
 if TYPE_CHECKING:
-    from ..device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 
 pytestmark = pytest.mark.models("core")
diff --git a/tests/click_tests/test_reset_slip39_advanced.py b/tests/click_tests/test_reset_slip39_advanced.py
index 12521bf8..d2f65196 100644
--- a/tests/click_tests/test_reset_slip39_advanced.py
+++ b/tests/click_tests/test_reset_slip39_advanced.py
@@ -19,13 +19,13 @@ from typing import TYPE_CHECKING
 import pytest
 
 from trezorlib import device, messages
+from trezorlib.testing import translations as TR
 
-from .. import translations as TR
 from ..common import EXTERNAL_ENTROPY, MOCK_GET_ENTROPY, generate_entropy
 from . import reset
 
 if TYPE_CHECKING:
-    from ..device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 
 pytestmark = pytest.mark.models("core")
diff --git a/tests/click_tests/test_reset_slip39_basic.py b/tests/click_tests/test_reset_slip39_basic.py
index d8f48767..3a156298 100644
--- a/tests/click_tests/test_reset_slip39_basic.py
+++ b/tests/click_tests/test_reset_slip39_basic.py
@@ -19,13 +19,13 @@ from typing import TYPE_CHECKING
 import pytest
 
 from trezorlib import device, messages
+from trezorlib.testing import translations as TR
 
-from .. import translations as TR
 from ..common import EXTERNAL_ENTROPY, MOCK_GET_ENTROPY, LayoutType, generate_entropy
 from . import reset
 
 if TYPE_CHECKING:
-    from ..device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 
 pytestmark = pytest.mark.models("core")
diff --git a/tests/click_tests/test_tutorial_caesar.py b/tests/click_tests/test_tutorial_caesar.py
index aa03518a..103a6cb9 100644
--- a/tests/click_tests/test_tutorial_caesar.py
+++ b/tests/click_tests/test_tutorial_caesar.py
@@ -21,13 +21,11 @@ import pytest
 
 from trezorlib import device
 from trezorlib.exceptions import Cancelled
-
-from .. import translations as TR
+from trezorlib.testing import translations as TR
 
 if TYPE_CHECKING:
     from trezorlib.debuglink import DebugLink
-
-    from ..device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 
 # Safe family only
diff --git a/tests/click_tests/test_tutorial_delizia.py b/tests/click_tests/test_tutorial_delizia.py
index 3bcd2df6..49695528 100644
--- a/tests/click_tests/test_tutorial_delizia.py
+++ b/tests/click_tests/test_tutorial_delizia.py
@@ -19,11 +19,10 @@ from typing import TYPE_CHECKING
 import pytest
 
 from trezorlib import device
-
-from .. import translations as TR
+from trezorlib.testing import translations as TR
 
 if TYPE_CHECKING:
-    from ..device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 
 # Trezor Safe 5 only
diff --git a/tests/click_tests/test_tutorial_eckhart.py b/tests/click_tests/test_tutorial_eckhart.py
index 37f61c23..513e801a 100644
--- a/tests/click_tests/test_tutorial_eckhart.py
+++ b/tests/click_tests/test_tutorial_eckhart.py
@@ -20,11 +20,10 @@ import pytest
 
 from trezorlib import device, exceptions
 from trezorlib.debuglink import DebugLink
-
-from .. import translations as TR
+from trezorlib.testing import translations as TR
 
 if TYPE_CHECKING:
-    from ..device_handler import BackgroundDeviceHandler
+    from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
 
 # Trezor Safe 7 only
diff --git a/tests/common.py b/tests/common.py
index be9ab1f8..a532a025 100644
--- a/tests/common.py
+++ b/tests/common.py
@@ -36,10 +36,6 @@ if TYPE_CHECKING:
     from trezorlib.debuglink import DebugLink
     from trezorlib.messages import ButtonRequest
 
-PRIVATE_KEYS_DEV = [byte * 32 for byte in (b"\xdd", b"\xde", b"\xdf")]
-
-BRGeneratorType = Generator[None, messages.ButtonRequest, None]
-
 
 # fmt: off
 #                1      2     3    4      5      6      7     8      9    10    11    12
@@ -386,44 +382,5 @@ def get_test_address(session: "Session") -> str:
     return btc.get_address(session, "Testnet", TEST_ADDRESS_N)
 
 
-def compact_size(n: int) -> bytes:
-    if n < 253:
-        return n.to_bytes(1, "little")
-    elif n < 0x1_0000:
-        return bytes([253]) + n.to_bytes(2, "little")
-    elif n < 0x1_0000_0000:
-        return bytes([254]) + n.to_bytes(4, "little")
-    else:
-        return bytes([255]) + n.to_bytes(8, "little")
-
-
-def get_text_possible_pagination(debug: "DebugLink", br: messages.ButtonRequest) -> str:
-    text = debug.read_layout().text_content()
-    if br.pages is not None:
-        for _ in range(br.pages - 1):
-            if debug.layout_type is LayoutType.Eckhart:
-                debug.click(debug.screen_buttons.ok())
-            else:
-                debug.swipe_up()
-            text += " "
-            text += debug.read_layout().text_content()
-    return text
-
-
-def swipe_if_necessary(
-    debug: "DebugLink", br_code: messages.ButtonRequestType | None = None
-) -> BRGeneratorType:
-    br = yield
-    if br_code is not None:
-        assert br.code == br_code
-    swipe_till_the_end(debug, br)
-
-
-def swipe_till_the_end(debug: "DebugLink", br: messages.ButtonRequest) -> None:
-    if br.pages is not None:
-        for _ in range(br.pages - 1):
-            debug.swipe_up()
-
-
 def is_core(session: Client | Session) -> bool:
     return session.model is not models.T1B1
diff --git a/tests/conftest.py b/tests/conftest.py
index c6dea67b..b7128491 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -39,8 +39,10 @@ from trezorlib.transport.ble import BleTransport
 # so that we see details of failed asserts from this module
 pytest.register_assert_rewrite("tests.common")
 
-from . import translations, ui_tests
-from .device_handler import BackgroundDeviceHandler
+from trezorlib.testing import translations
+from trezorlib.testing.device_handler import BackgroundDeviceHandler
+
+from . import ui_tests
 from .emulators import EmulatorWrapper
 
 if t.TYPE_CHECKING:
diff --git a/tests/definitions.py b/tests/definitions.py
index a18996cd..a83350ab 100644
--- a/tests/definitions.py
+++ b/tests/definitions.py
@@ -5,8 +5,7 @@ import typing as t
 from hashlib import sha256
 
 from trezorlib import cosi, definitions, messages, protobuf
-
-from .common import PRIVATE_KEYS_DEV
+from trezorlib.testing.common import PRIVATE_KEYS_DEV
 
 
 def make_eth_network(
diff --git a/tests/device_handler.py b/tests/device_handler.py
deleted file mode 100644
index 3f25f9af..00000000
--- a/tests/device_handler.py
+++ /dev/null
@@ -1,149 +0,0 @@
-from __future__ import annotations
-
-import typing as t
-from concurrent.futures import ThreadPoolExecutor
-
-import typing_extensions as tx
-
-from trezorlib.messages import DebugWaitType
-from trezorlib.transport import udp
-
-if t.TYPE_CHECKING:
-    from trezorlib._internal.emulator import Emulator
-    from trezorlib.client import Session
-    from trezorlib.debuglink import DebugLink
-    from trezorlib.debuglink import TrezorTestContext as Client
-    from trezorlib.messages import Features
-
-    P = tx.ParamSpec("P")
-
-
-udp.SOCKET_TIMEOUT = 0.1
-
-
-class NullUI:
-    @staticmethod
-    def clear(*args, **kwargs):
-        pass
-
-    @staticmethod
-    def button_request(code):
-        pass
-
-    @staticmethod
-    def get_pin(code=None):
-        raise NotImplementedError("NullUI should not be used with T1")
-
-
-class BackgroundDeviceHandler:
-    _pool = ThreadPoolExecutor()
-
-    def __init__(self, client: "Client", nowait: bool = False) -> None:
-        self._configure_client(client)
-        self.task = None
-        self.nowait = nowait
-
-    def _configure_client(self, client: "Client") -> None:
-        self.client = client
-        self.client.ui = NullUI  # type: ignore [NullUI is OK UI]
-        self.client.app.button_callback = self.client.ui.button_request
-        self.client.debug.input_wait_type = DebugWaitType.CURRENT_LAYOUT
-
-    def get_session(self, *args, **kwargs):
-        if self.task is not None:
-            raise RuntimeError("Wait for previous task first")
-
-        with self.debuglink().wait_for_layout_change():
-            self.task = self._pool.submit(self.client.get_session, *args, **kwargs)
-
-    def run_with_session(
-        self,
-        function: t.Callable[tx.Concatenate["Session", P], t.Any],
-        seedless: bool = False,
-        *args: P.args,
-        **kwargs: P.kwargs,
-    ) -> None:
-        """Runs some function that interacts with a device.
-
-        Makes sure the UI is updated before returning.
-        """
-        if self.task is not None:
-            raise RuntimeError("Wait for previous task first")
-
-        def task_function(*args, **kwargs):
-            if seedless:
-                session = self.client.get_seedless_session()
-            else:
-                session = self.client.get_session()
-            return function(session, *args, **kwargs)
-
-        # wait for the first UI change triggered by the task running in the background
-        with self.debuglink().wait_for_layout_change():
-            self.task = self._pool.submit(task_function, *args, **kwargs)
-
-    def run_with_provided_session(
-        self,
-        session,
-        function: t.Callable[tx.Concatenate["Session", P], t.Any],
-        *args: P.args,
-        **kwargs: P.kwargs,
-    ) -> None:
-        """Runs some function that interacts with a device.
-
-        Makes sure the UI is updated before returning.
-        """
-        if self.task is not None:
-            raise RuntimeError("Wait for previous task first")
-
-        # wait for the first UI change triggered by the task running in the background
-        with self.debuglink().wait_for_layout_change():
-            self.task = self._pool.submit(function, session, *args, **kwargs)
-
-    def kill_task(self) -> None:
-        if self.task is not None:
-            # Force close the transport, which should raise an exception in a client
-            # waiting on IO. Does not work over Bridge, because bridge doesn't have
-            # a close() method.
-            self.client.transport.close()
-            try:
-                self.task.result(timeout=1)
-            except Exception:
-                pass
-        self.task = None
-
-    def restart(self, emulator: "Emulator") -> None:
-        # TODO handle actual restart as well
-        self.kill_task()
-        emulator.restart()
-        self._configure_client(emulator.client)  # type: ignore [client cannot be None]
-
-    def result(self, timeout: float | None = None) -> t.Any:
-        if self.task is None:
-            raise RuntimeError("No task running")
-        try:
-            return self.task.result(timeout=timeout)
-        finally:
-            self.task = None
-
-    def features(self) -> "Features":
-        if self.task is not None:
-            raise RuntimeError("Cannot query features while task is running")
-        self.client.refresh_features()
-        return self.client.features
-
-    def debuglink(self) -> "DebugLink":
-        return self.client.debug
-
-    def check_finalize(self) -> bool:
-        if self.task is not None:
-            self.kill_task()
-            return False
-        return True
-
-    def __enter__(self) -> "BackgroundDeviceHandler":
-        return self
-
-    def __exit__(self, exc_type, exc_value, traceback) -> None:
-        finalized_ok = self.check_finalize()
-        if exc_type is None and not finalized_ok:
-            raise RuntimeError("Exit while task is unfinished")
diff --git a/tests/device_tests/evolu/common.py b/tests/device_tests/evolu/common.py
index 38de6d69..141c854d 100644
--- a/tests/device_tests/evolu/common.py
+++ b/tests/device_tests/evolu/common.py
@@ -8,10 +8,9 @@ from trezorlib import evolu
 from trezorlib.debuglink import DebugSession as Session
 from trezorlib.debuglink import TrezorTestContext as Client
 from trezorlib.messages import EvoluDelegatedIdentityKey, ThpCredentialResponse
+from trezorlib.testing.common import compact_size
 from trezorlib.thp import curve25519
 
-from ...common import compact_size
-
 TEST_randomness = os.urandom(32)
 TEST_host_static_private_key = curve25519.get_private_key(TEST_randomness)
 TEST_host_static_public_key = curve25519.get_public_key(TEST_host_static_private_key)
diff --git a/tests/device_tests/evolu/test_sign_registration.py b/tests/device_tests/evolu/test_sign_registration.py
index c5bcfa2d..2471d334 100644
--- a/tests/device_tests/evolu/test_sign_registration.py
+++ b/tests/device_tests/evolu/test_sign_registration.py
@@ -4,8 +4,8 @@ from ecdsa import NIST256p, SigningKey, VerifyingKey
 from trezorlib import evolu
 from trezorlib.debuglink import TrezorTestContext as Client
 from trezorlib.exceptions import TrezorFailure
+from trezorlib.testing.common import compact_size
 
-from ...common import compact_size
 from ..certificate import check_signature_optiga
 from .common import get_delegated_identity_key, get_invalid_proof, get_proof
 
diff --git a/tests/device_tests/misc/test_msg_enablelabeling.py b/tests/device_tests/misc/test_msg_enablelabeling.py
index cd883e7a..8b2f4d0a 100644
--- a/tests/device_tests/misc/test_msg_enablelabeling.py
+++ b/tests/device_tests/misc/test_msg_enablelabeling.py
@@ -18,8 +18,8 @@ import pytest
 
 from trezorlib import misc
 from trezorlib.debuglink import TrezorTestContext as Client
+from trezorlib.testing import translations as TR
 
-from ... import translations as TR
 from ...common import MNEMONIC12
 
 
diff --git a/tests/device_tests/payment_req.py b/tests/device_tests/payment_req.py
index e1c14435..6b33735a 100644
--- a/tests/device_tests/payment_req.py
+++ b/tests/device_tests/payment_req.py
@@ -5,8 +5,7 @@ from ecdsa import NIST256p, SigningKey
 
 from trezorlib import messages
 from trezorlib.client import Session
-
-from ..common import compact_size
+from trezorlib.testing.common import compact_size
 
 SLIP44_ID_UNDEFINED = 0xFFFF_FFFF
 
diff --git a/tests/device_tests/reset_recovery/test_recovery_restart.py b/tests/device_tests/reset_recovery/test_recovery_restart.py
index b0e7dc9a..22b2e390 100644
--- a/tests/device_tests/reset_recovery/test_recovery_restart.py
+++ b/tests/device_tests/reset_recovery/test_recovery_restart.py
@@ -16,12 +16,12 @@
 
 import pytest
 
-from tests import translations as TR
 from trezorlib import device
 from trezorlib.debuglink import DebugSession as Session
 from trezorlib.debuglink import LayoutType
 from trezorlib.exceptions import Cancelled
 from trezorlib.messages import BackupMethod, Capability, RecoveryStatus
+from trezorlib.testing import translations as TR
 
 from ...input_flows import RecoveryFlow
 
diff --git a/tests/device_tests/reset_recovery/test_reset_bip39_t1.py b/tests/device_tests/reset_recovery/test_reset_bip39_t1.py
index 44ad3c61..934c9e16 100644
--- a/tests/device_tests/reset_recovery/test_reset_bip39_t1.py
+++ b/tests/device_tests/reset_recovery/test_reset_bip39_t1.py
@@ -25,14 +25,10 @@ from trezorlib.btc import get_public_node
 from trezorlib.debuglink import DebugLink
 from trezorlib.debuglink import DebugSession as Session
 from trezorlib.debuglink import TrezorTestContext
+from trezorlib.testing.common import BRGeneratorType
 from trezorlib.tools import parse_path
 
-from ...common import (
-    EXTERNAL_ENTROPY,
-    MOCK_GET_ENTROPY,
-    BRGeneratorType,
-    generate_entropy,
-)
+from ...common import EXTERNAL_ENTROPY, MOCK_GET_ENTROPY, generate_entropy
 
 pytestmark = pytest.mark.models("legacy")
 
diff --git a/tests/device_tests/reset_recovery/test_reset_recovery_bip39.py b/tests/device_tests/reset_recovery/test_reset_recovery_bip39.py
index a1e11bcc..3e23b5ff 100644
--- a/tests/device_tests/reset_recovery/test_reset_recovery_bip39.py
+++ b/tests/device_tests/reset_recovery/test_reset_recovery_bip39.py
@@ -20,11 +20,11 @@ from trezorlib import btc, device, messages
 from trezorlib.debuglink import DebugSession as Session
 from trezorlib.debuglink import TrezorTestContext as Client
 from trezorlib.messages import BackupMethod, BackupType
+from trezorlib.testing.translations import set_language
 from trezorlib.tools import parse_path
 
 from ...common import MOCK_GET_ENTROPY
 from ...input_flows import InputFlowBip39Recovery, InputFlowBip39ResetBackup
-from ...translations import set_language
 
 
 @pytest.mark.models("core")
diff --git a/tests/device_tests/reset_recovery/test_reset_recovery_slip39_advanced.py b/tests/device_tests/reset_recovery/test_reset_recovery_slip39_advanced.py
index cd8fe808..7e9f17e8 100644
--- a/tests/device_tests/reset_recovery/test_reset_recovery_slip39_advanced.py
+++ b/tests/device_tests/reset_recovery/test_reset_recovery_slip39_advanced.py
@@ -20,6 +20,7 @@ from trezorlib import btc, device, messages
 from trezorlib.debuglink import DebugSession as Session
 from trezorlib.debuglink import TrezorTestContext as Client
 from trezorlib.messages import BackupMethod, BackupType
+from trezorlib.testing.translations import set_language
 from trezorlib.tools import parse_path
 
 from ...common import MOCK_GET_ENTROPY
@@ -27,7 +28,6 @@ from ...input_flows import (
     InputFlowSlip39AdvancedRecovery,
     InputFlowSlip39AdvancedResetRecovery,
 )
-from ...translations import set_language
 
 
 @pytest.mark.models("core")
diff --git a/tests/device_tests/reset_recovery/test_reset_recovery_slip39_basic.py b/tests/device_tests/reset_recovery/test_reset_recovery_slip39_basic.py
index aad2e0cc..f817d108 100644
--- a/tests/device_tests/reset_recovery/test_reset_recovery_slip39_basic.py
+++ b/tests/device_tests/reset_recovery/test_reset_recovery_slip39_basic.py
@@ -23,6 +23,7 @@ from trezorlib import btc, device, messages
 from trezorlib.debuglink import DebugSession as Session
 from trezorlib.debuglink import TrezorTestContext as Client
 from trezorlib.messages import BackupMethod, BackupType
+from trezorlib.testing.translations import set_language
 from trezorlib.tools import parse_path
 
 from ...common import MOCK_GET_ENTROPY
@@ -30,7 +31,6 @@ from ...input_flows import (
     InputFlowSlip39BasicRecovery,
     InputFlowSlip39BasicResetRecovery,
 )
-from ...translations import set_language
 
 
 @pytest.mark.models("core")
diff --git a/tests/device_tests/test_authenticate_device.py b/tests/device_tests/test_authenticate_device.py
index 5d1e2d94..a5c8ac9e 100644
--- a/tests/device_tests/test_authenticate_device.py
+++ b/tests/device_tests/test_authenticate_device.py
@@ -2,8 +2,8 @@ import pytest
 
 from trezorlib import device, exceptions, messages
 from trezorlib.debuglink import DebugSession as Session
+from trezorlib.testing.common import compact_size
 
-from ..common import compact_size
 from .certificate import check_signature_optiga, check_signature_tropic
 
 # The tests below require Optiga (and some require Tropic)
diff --git a/tests/device_tests/test_language.py b/tests/device_tests/test_language.py
index f3d463b1..589b02dc 100644
--- a/tests/device_tests/test_language.py
+++ b/tests/device_tests/test_language.py
@@ -26,8 +26,7 @@ from trezorlib._internal import translations
 from trezorlib.debuglink import DebugSession as Session
 from trezorlib.debuglink import TrezorTestContext as Client
 from trezorlib.debuglink import message_filters
-
-from ..translations import (
+from trezorlib.testing.translations import (
     LANGUAGES,
     build_and_sign_blob,
     get_lang_json,
diff --git a/tests/device_tests/test_repeated_backup.py b/tests/device_tests/test_repeated_backup.py
index fecdf4c7..f1a70923 100644
--- a/tests/device_tests/test_repeated_backup.py
+++ b/tests/device_tests/test_repeated_backup.py
@@ -20,8 +20,8 @@ import pytest
 from trezorlib import device, exceptions, messages
 from trezorlib.debuglink import DebugSession as Session
 from trezorlib.exceptions import Cancelled, TrezorFailure
+from trezorlib.testing import translations as TR
 
-from .. import translations as TR
 from ..common import (
     MNEMONIC_SLIP39_SINGLE_EXT_20,
     TEST_ADDRESS_N,
diff --git a/tests/device_tests/test_sdcard.py b/tests/device_tests/test_sdcard.py
index 34409efc..23843ec2 100644
--- a/tests/device_tests/test_sdcard.py
+++ b/tests/device_tests/test_sdcard.py
@@ -26,7 +26,7 @@ from trezorlib.messages import SdProtectOperationType as Op
 
 B = messages.ButtonRequestType
 
-from .. import translations as TR
+from trezorlib.testing import translations as TR
 
 PIN = "1234"
 
diff --git a/tests/device_tests/test_session_id_and_passphrase.py b/tests/device_tests/test_session_id_and_passphrase.py
index 77caace4..b5cd6da5 100644
--- a/tests/device_tests/test_session_id_and_passphrase.py
+++ b/tests/device_tests/test_session_id_and_passphrase.py
@@ -25,10 +25,9 @@ from trezorlib.debuglink import LayoutType, TrezorTestContext
 from trezorlib.exceptions import TrezorFailure
 from trezorlib.messages import FailureType, SafetyCheckLevel
 from trezorlib.protocol_v1 import SessionV1, TrezorClientV1
+from trezorlib.testing import translations as TR
 from trezorlib.tools import parse_path
 
-from .. import translations as TR
-
 pytestmark = pytest.mark.protocol("v1")
 
 XPUB_PASSPHRASES = {
diff --git a/tests/input_flows.py b/tests/input_flows.py
index 910bd923..b0a624fb 100644
--- a/tests/input_flows.py
+++ b/tests/input_flows.py
@@ -22,17 +22,19 @@ from trezorlib.debuglink import DebugLink, DebugSession, LayoutContent, LayoutTy
 from trezorlib.debuglink import TrezorTestContext as Client
 from trezorlib.debuglink import multipage_content
 from trezorlib.exceptions import TrezorFailure
+from trezorlib.testing import translations as TR
+from trezorlib.testing.common import (
+    BRGeneratorType,
+    get_text_possible_pagination,
+    swipe_if_necessary,
+)
 
-from . import translations as TR
 from .common import (
-    BRGeneratorType,
     check_pin_backoff_time,
     click_info_button_bolt,
     click_info_button_delizia_eckhart,
     click_through,
-    get_text_possible_pagination,
     read_and_confirm_mnemonic,
-    swipe_if_necessary,
 )
 from .input_flows_helpers import (
     BackupFlow,
diff --git a/tests/input_flows_helpers.py b/tests/input_flows_helpers.py
index c42981e0..1aa2492d 100644
--- a/tests/input_flows_helpers.py
+++ b/tests/input_flows_helpers.py
@@ -3,9 +3,8 @@ import typing as t
 from trezorlib import messages
 from trezorlib.debuglink import DebugLink, LayoutType
 from trezorlib.debuglink import TrezorTestContext as Client
-
-from . import translations as TR
-from .common import BRGeneratorType, get_text_possible_pagination
+from trezorlib.testing import translations as TR
+from trezorlib.testing.common import BRGeneratorType, get_text_possible_pagination
 
 B = messages.ButtonRequestType
 
diff --git a/tests/persistence_tests/test_shamir_persistence.py b/tests/persistence_tests/test_shamir_persistence.py
index 3d75838b..fe0dab7c 100644
--- a/tests/persistence_tests/test_shamir_persistence.py
+++ b/tests/persistence_tests/test_shamir_persistence.py
@@ -19,11 +19,11 @@ import pytest
 from trezorlib import device, messages
 from trezorlib.debuglink import DebugLink, LayoutType
 from trezorlib.messages import RecoveryStatus
+from trezorlib.testing import translations as TR
+from trezorlib.testing.device_handler import BackgroundDeviceHandler
 
-from .. import translations as TR
 from ..click_tests import common, recovery
 from ..common import MNEMONIC_SLIP39_ADVANCED_20, MNEMONIC_SLIP39_BASIC_20_3of6
-from ..device_handler import BackgroundDeviceHandler
 from ..emulators import Emulator
 from ..upgrade_tests import core_only
 
diff --git a/tests/translations.py b/tests/translations.py
deleted file mode 100644
index 53bfc081..00000000
--- a/tests/translations.py
+++ /dev/null
@@ -1,175 +0,0 @@
-from __future__ import annotations
-
-import json
-import re
-import threading
-import typing as t
-import warnings
-from hashlib import sha256
-from pathlib import Path
-
-from trezorlib import cosi, device, models
-from trezorlib._internal import translations
-from trezorlib.debuglink import DebugSession, LayoutType
-
-from . import common
-
-HERE = Path(__file__).resolve().parent
-ROOT = HERE.parent
-
-TRANSLATIONS_DIR = ROOT / "core" / "translations"
-FONTS_DIR = TRANSLATIONS_DIR / "fonts"
-ORDER_FILE = TRANSLATIONS_DIR / "order.json"
-
-LANGUAGES = [file.stem for file in TRANSLATIONS_DIR.glob("??.json")]
-
-_CURRENT_TRANSLATION = threading.local()
-
-
-def prepare_blob(
-    lang_or_def: translations.JsonDef | Path | str,
-    model: models.TrezorModel,
-    version: translations.VersionTuple | tuple[int, int, int] | None = None,
-) -> translations.TranslationsBlob:
-    order = translations.order_from_json(json.loads(ORDER_FILE.read_text()))
-    if isinstance(lang_or_def, str):
-        lang_or_def = get_lang_json(lang_or_def)
-    if isinstance(lang_or_def, Path):
-        lang_or_def = t.cast(translations.JsonDef, json.loads(lang_or_def.read_text()))
-
-    # generate raw blob
-    if version is None:
-        version = translations.version_from_json(lang_or_def["header"]["version"])
-    elif len(version) == 3:
-        # version coming from client object does not have build item
-        version = *version, 0
-    return translations.blob_from_defs(lang_or_def, order, model, version, FONTS_DIR)
-
-
-def sign_blob(blob: translations.TranslationsBlob) -> bytes:
-    # build 0-item Merkle proof
-    digest = sha256(b"\x00" + blob.header_bytes).digest()
-    signature = cosi.sign_with_privkeys(digest, common.PRIVATE_KEYS_DEV)
-    blob.proof = translations.Proof(
-        merkle_proof=[],
-        sigmask=0b111,
-        signature=signature,
-    )
-    return blob.build()
-
-
-def build_and_sign_blob(
-    lang_or_def: translations.JsonDef | Path | str,
-    session: DebugSession,
-) -> bytes:
-    blob = prepare_blob(lang_or_def, session.model, session.version)
-    return sign_blob(blob)
-
-
-def set_language(session: DebugSession, lang: str, *, force: bool = False):
-    if lang.startswith("en"):
-        language_data = b""
-    else:
-        language_data = build_and_sign_blob(lang, session)
-    with session.test_ctx:
-        if not session.features.language.startswith(lang) or force:
-            device.change_language(session, language_data)  # type: ignore
-    _CURRENT_TRANSLATION.LAYOUT = session.layout_type
-    _CURRENT_TRANSLATION.TR = TRANSLATIONS[lang]
-
-
-def get_language() -> str:
-    for lang in LANGUAGES:
-        if _CURRENT_TRANSLATION.TR == TRANSLATIONS[lang]:
-            return lang
-    return "en"
-
-
-def get_lang_json(lang: str) -> translations.JsonDef:
-    assert lang in LANGUAGES
-    lang_json = json.loads((TRANSLATIONS_DIR / f"{lang}.json").read_text())
-    if (fonts_safe3 := lang_json.get("fonts", {}).get("##Safe3")) is not None:
-        lang_json["fonts"]["T2B1"] = fonts_safe3
-        lang_json["fonts"]["T3B1"] = fonts_safe3
-    return lang_json
-
-
-class Translation:
-    FORMAT_STR_RE = re.compile(r"\\{\d+\\}")
-
-    def __init__(self, lang: str) -> None:
-        self.lang = lang
-        self.lang_json = get_lang_json(lang)
-
-    @property
-    def translations(self) -> dict[str, str | dict[str, str]]:
-        return self.lang_json["translations"]
-
-    def _translate_raw(self, key: str, _stacklevel: int = 0) -> str:
-        tr = self.translations.get(key)
-        # Emulate firmware behaviour: English strings are used instead of missing & empty strings
-        if tr:
-            # Handle layout-specific translations
-            if isinstance(tr, dict) and hasattr(_CURRENT_TRANSLATION, "LAYOUT"):
-                # Try to get translation for current layout
-                layout_name = _CURRENT_TRANSLATION.LAYOUT.name
-                if layout_name in tr:
-                    return tr[layout_name]
-                # Fall back to any available translation if no match for current layout
-                return next(iter(tr.values()))
-            elif isinstance(tr, str):
-                return tr
-            else:
-                raise ValueError(f"Invalid translation value for key '{key}'")
-        if self.lang != "en":
-            # check if the key exists in English first
-            retval = TRANSLATIONS["en"]._translate_raw(key)
-            # if not, a KeyError was raised so we fall through.
-            # otherwise, warn that the key is untranslated in target language.
-            warnings.warn(
-                f"Translation key '{key}' not found in '{self.lang}' translation file",
-                stacklevel=_stacklevel + 2,
-            )
-            return retval
-        raise KeyError(key)
-
-    def translate(self, key: str, _stacklevel: int = 0) -> str:
-        return self._translate_raw(key, _stacklevel=_stacklevel + 1).strip()
-
-    def as_regexp(self, key: str, _stacklevel: int = 0) -> re.Pattern:
-        tr = self.translate(key, _stacklevel=_stacklevel + 1)
-        re_safe = re.escape(tr)
-        return re.compile(self.FORMAT_STR_RE.sub(r".*?", re_safe))
-
-    def format(self, key: str, *args, **kwargs) -> str:
-        tr = self.translate(key)
-        try:
-            return tr.format(*args, **kwargs)
-        except (KeyError, IndexError) as e:
-            raise ValueError(
-                f"Failed to format translation '{key}' with args={args}, kwargs={kwargs}: {e}"
-            ) from e
-
-
-TRANSLATIONS = {lang: Translation(lang) for lang in LANGUAGES}
-_CURRENT_TRANSLATION.TR = TRANSLATIONS["en"]
-_CURRENT_TRANSLATION.LAYOUT = LayoutType.Bolt
-
-
-def translate(key: str, _stacklevel: int = 0) -> str:
-    return _CURRENT_TRANSLATION.TR.translate(key, _stacklevel=_stacklevel + 1)
-
-
-def regexp(key: str) -> re.Pattern:
-    return _CURRENT_TRANSLATION.TR.as_regexp(key, _stacklevel=1)
-
-
-def format(key: str, *args, **kwargs) -> str:
-    return _CURRENT_TRANSLATION.TR.format(key, *args, **kwargs)
-
-
-def __getattr__(key: str) -> str:
-    try:
-        return translate(key, _stacklevel=1)
-    except KeyError as e:
-        raise AttributeError(f"Translation key '{key}' not found") from e
diff --git a/tests/upgrade_tests/test_firmware_upgrades.py b/tests/upgrade_tests/test_firmware_upgrades.py
index d7bd3174..132b434d 100644
--- a/tests/upgrade_tests/test_firmware_upgrades.py
+++ b/tests/upgrade_tests/test_firmware_upgrades.py
@@ -33,11 +33,11 @@ from trezorlib.messages import (
     RecoveryStatus,
     Success,
 )
+from trezorlib.testing.device_handler import BackgroundDeviceHandler
 from trezorlib.tools import H_, parse_path
 
 from ..click_tests import recovery
 from ..common import MNEMONIC_SLIP39_BASIC_20_3of6, MNEMONIC_SLIP39_BASIC_20_3of6_SECRET
-from ..device_handler import BackgroundDeviceHandler
 from ..emulators import EmulatorWrapper
 from ..input_flows import InputFlowSlip39BasicBackup
 from . import ALL_TAGS, for_all, for_tags, recovery_old, version_from_tag