diff --git a/homeassistant/components/tts/__init__.py b/homeassistant/components/tts/__init__.py
index 515dc25899a..abe3d29c607 100644
--- a/homeassistant/components/tts/__init__.py
+++ b/homeassistant/components/tts/__init__.py
@@ -9,6 +9,7 @@ import io
 import logging
 import mimetypes
 import os
+from pathlib import Path
 import re
 from typing import TYPE_CHECKING, Optional, cast
 
@@ -39,10 +40,11 @@ import homeassistant.helpers.config_validation as cv
 from homeassistant.helpers.network import get_url
 from homeassistant.helpers.service import async_set_service_schema
 from homeassistant.helpers.typing import ConfigType, DiscoveryInfoType
-from homeassistant.loader import async_get_integration
 from homeassistant.setup import async_prepare_setup_platform
 from homeassistant.util.yaml import load_yaml
 
+from .const import DOMAIN
+
 # mypy: allow-untyped-defs, no-check-untyped-defs
 
 _LOGGER = logging.getLogger(__name__)
@@ -69,7 +71,6 @@ CONF_FIELDS = "fields"
 DEFAULT_CACHE = True
 DEFAULT_CACHE_DIR = "tts"
 DEFAULT_TIME_MEMORY = 300
-DOMAIN = "tts"
 
 MEM_CACHE_FILENAME = "filename"
 MEM_CACHE_VOICE = "voice"
@@ -135,12 +136,12 @@ async def async_setup(hass: HomeAssistant, config: ConfigType) -> bool:
         _LOGGER.exception("Error on cache init")
         return False
 
+    hass.data[DOMAIN] = tts
     hass.http.register_view(TextToSpeechView(tts))
     hass.http.register_view(TextToSpeechUrlView(tts))
 
     # Load service descriptions from tts/services.yaml
-    integration = await async_get_integration(hass, DOMAIN)
-    services_yaml = integration.file_path / "services.yaml"
+    services_yaml = Path(__file__).parent / "services.yaml"
     services_dict = cast(
         dict, await hass.async_add_executor_job(load_yaml, str(services_yaml))
     )
@@ -343,7 +344,11 @@ class SpeechManager:
 
         This method is a coroutine.
         """
-        provider = self.providers[engine]
+        provider = self.providers.get(engine)
+
+        if provider is None:
+            raise HomeAssistantError(f"Provider {engine} not found")
+
         msg_hash = hashlib.sha1(bytes(message, "utf-8")).hexdigest()
         use_cache = cache if cache is not None else self.use_cache
 
diff --git a/homeassistant/components/tts/const.py b/homeassistant/components/tts/const.py
new file mode 100644
index 00000000000..492e995b87f
--- /dev/null
+++ b/homeassistant/components/tts/const.py
@@ -0,0 +1,3 @@
+"""Text-to-speech constants."""
+
+DOMAIN = "tts"
diff --git a/homeassistant/components/tts/media_source.py b/homeassistant/components/tts/media_source.py
new file mode 100644
index 00000000000..2398a6203ad
--- /dev/null
+++ b/homeassistant/components/tts/media_source.py
@@ -0,0 +1,109 @@
+"""Text-to-speech media source."""
+from __future__ import annotations
+
+import mimetypes
+from typing import TYPE_CHECKING
+
+from yarl import URL
+
+from homeassistant.components.media_player.const import MEDIA_CLASS_APP
+from homeassistant.components.media_player.errors import BrowseError
+from homeassistant.components.media_source.error import Unresolvable
+from homeassistant.components.media_source.models import (
+    BrowseMediaSource,
+    MediaSource,
+    MediaSourceItem,
+    PlayMedia,
+)
+from homeassistant.core import HomeAssistant, callback
+from homeassistant.exceptions import HomeAssistantError
+
+from .const import DOMAIN
+
+if TYPE_CHECKING:
+    from . import SpeechManager
+
+
+async def async_get_media_source(hass: HomeAssistant) -> TTSMediaSource:
+    """Set up tts media source."""
+    return TTSMediaSource(hass)
+
+
+class TTSMediaSource(MediaSource):
+    """Provide text-to-speech providers as media sources."""
+
+    name: str = "Text to Speech"
+
+    def __init__(self, hass: HomeAssistant) -> None:
+        """Initialize TTSMediaSource."""
+        super().__init__(DOMAIN)
+        self.hass = hass
+
+    async def async_resolve_media(self, item: MediaSourceItem) -> PlayMedia:
+        """Resolve media to a url."""
+        parsed = URL(item.identifier)
+        if "message" not in parsed.query:
+            raise Unresolvable("No message specified.")
+
+        options = dict(parsed.query)
+        kwargs = {
+            "engine": parsed.name,
+            "message": options.pop("message"),
+            "language": options.pop("language", None),
+            "options": options,
+        }
+
+        manager: SpeechManager = self.hass.data[DOMAIN]
+
+        try:
+            url = await manager.async_get_url_path(**kwargs)  # type: ignore
+        except HomeAssistantError as err:
+            raise Unresolvable(str(err)) from err
+
+        mime_type = mimetypes.guess_type(url)[0] or "audio/mpeg"
+
+        return PlayMedia(url, mime_type)
+
+    async def async_browse_media(
+        self,
+        item: MediaSourceItem,
+    ) -> BrowseMediaSource:
+        """Return media."""
+        if item.identifier:
+            provider, _, _ = item.identifier.partition("?")
+            return self._provider_item(provider)
+
+        # Root. List providers.
+        manager: SpeechManager = self.hass.data[DOMAIN]
+        children = [self._provider_item(provider) for provider in manager.providers]
+        return BrowseMediaSource(
+            domain=DOMAIN,
+            identifier=None,
+            media_class=MEDIA_CLASS_APP,
+            media_content_type="",
+            title=self.name,
+            can_play=False,
+            can_expand=True,
+            children_media_class=MEDIA_CLASS_APP,
+            children=children,
+        )
+
+    @callback
+    def _provider_item(self, provider_domain: str) -> BrowseMediaSource:
+        """Return provider item."""
+        manager: SpeechManager = self.hass.data[DOMAIN]
+        provider = manager.providers.get(provider_domain)
+
+        if provider is None:
+            raise BrowseError("Unknown provider")
+
+        return BrowseMediaSource(
+            domain=DOMAIN,
+            identifier=provider_domain,
+            media_class=MEDIA_CLASS_APP,
+            media_content_type="provider",
+            title=provider.name,
+            thumbnail=f"https://brands.home-assistant.io/_/{provider_domain}/logo.png",
+            can_play=False,
+            can_expand=True,
+        )
diff --git a/tests/components/google_translate/test_tts.py b/tests/components/google_translate/test_tts.py
index 5690591ccd2..9e09ebb9ff2 100644
--- a/tests/components/google_translate/test_tts.py
+++ b/tests/components/google_translate/test_tts.py
@@ -16,7 +16,7 @@ from homeassistant.config import async_process_ha_core_config
 from homeassistant.setup import async_setup_component
 
 from tests.common import async_mock_service
-from tests.components.tts.test_init import mutagen_mock  # noqa: F401
+from tests.components.tts.conftest import mutagen_mock  # noqa: F401
 
 
 @pytest.fixture(autouse=True)
diff --git a/tests/components/tts/conftest.py b/tests/components/tts/conftest.py
index 3580880fedb..6d995978391 100644
--- a/tests/components/tts/conftest.py
+++ b/tests/components/tts/conftest.py
@@ -2,9 +2,12 @@
 
 From http://doc.pytest.org/en/latest/example/simple.html#making-test-result-information-available-in-fixtures
 """
+from unittest.mock import patch
 
 import pytest
 
+from homeassistant.components.tts import _get_cache_files
+
 
 @pytest.hookimpl(tryfirst=True, hookwrapper=True)
 def pytest_runtest_makereport(item, call):
@@ -16,3 +19,55 @@ def pytest_runtest_makereport(item, call):
     # set a report attribute for each phase of a call, which can
     # be "setup", "call", "teardown"
     setattr(item, f"rep_{rep.when}", rep)
+
+
+@pytest.fixture(autouse=True)
+def mock_get_cache_files():
+    """Mock the list TTS cache function."""
+    with patch(
+        "homeassistant.components.tts._get_cache_files", return_value={}
+    ) as mock_cache_files:
+        yield mock_cache_files
+
+
+@pytest.fixture(autouse=True)
+def mock_init_cache_dir():
+    """Mock the TTS cache dir in memory."""
+    with patch(
+        "homeassistant.components.tts._init_tts_cache_dir",
+        side_effect=lambda hass, cache_dir: hass.config.path(cache_dir),
+    ) as mock_cache_dir:
+        yield mock_cache_dir
+
+
+@pytest.fixture
+def empty_cache_dir(tmp_path, mock_init_cache_dir, mock_get_cache_files, request):
+    """Mock the TTS cache dir with empty dir."""
+    mock_init_cache_dir.side_effect = None
+    mock_init_cache_dir.return_value = str(tmp_path)
+
+    # Restore original get cache files behavior, we're working with a real dir.
+    mock_get_cache_files.side_effect = _get_cache_files
+
+    yield tmp_path
+
+    if request.node.rep_call.passed:
+        return
+
+    # Print contents of dir if failed
+    print("Content of dir for", request.node.nodeid)
+    for fil in tmp_path.iterdir():
+        print(fil.relative_to(tmp_path))
+
+    # To show the log.
+    assert False
+
+
+@pytest.fixture(autouse=True)
+def mutagen_mock():
+    """Mock writing tags."""
+    with patch(
+        "homeassistant.components.tts.SpeechManager.write_tags",
+        side_effect=lambda *args: args[1],
+    ) as mock_write_tags:
+        yield mock_write_tags
diff --git a/tests/components/tts/test_init.py b/tests/components/tts/test_init.py
index 3cbc1f0da00..9f1cc849a1f 100644
--- a/tests/components/tts/test_init.py
+++ b/tests/components/tts/test_init.py
@@ -5,6 +5,7 @@ from unittest.mock import PropertyMock, patch
 import pytest
 import yarl
 
+from homeassistant.components import tts
 from homeassistant.components.demo.tts import DemoProvider
 from homeassistant.components.media_player.const import (
     ATTR_MEDIA_CONTENT_ID,
@@ -13,13 +14,13 @@ from homeassistant.components.media_player.const import (
     MEDIA_TYPE_MUSIC,
     SERVICE_PLAY_MEDIA,
 )
-import homeassistant.components.tts as tts
-from homeassistant.components.tts import _get_cache_files
 from homeassistant.config import async_process_ha_core_config
 from homeassistant.setup import async_setup_component
 
 from tests.common import assert_setup_component, async_mock_service
 
+ORIG_WRITE_TAGS = tts.SpeechManager.write_tags
+
 
 def relative_url(url):
     """Convert an absolute url to a relative one."""
@@ -32,58 +33,6 @@ def demo_provider():
     return DemoProvider("en")
 
 
-@pytest.fixture(autouse=True)
-def mock_get_cache_files():
-    """Mock the list TTS cache function."""
-    with patch(
-        "homeassistant.components.tts._get_cache_files", return_value={}
-    ) as mock_cache_files:
-        yield mock_cache_files
-
-
-@pytest.fixture(autouse=True)
-def mock_init_cache_dir():
-    """Mock the TTS cache dir in memory."""
-    with patch(
-        "homeassistant.components.tts._init_tts_cache_dir",
-        side_effect=lambda hass, cache_dir: hass.config.path(cache_dir),
-    ) as mock_cache_dir:
-        yield mock_cache_dir
-
-
-@pytest.fixture
-def empty_cache_dir(tmp_path, mock_init_cache_dir, mock_get_cache_files, request):
-    """Mock the TTS cache dir with empty dir."""
-    mock_init_cache_dir.side_effect = None
-    mock_init_cache_dir.return_value = str(tmp_path)
-
-    # Restore original get cache files behavior, we're working with a real dir.
-    mock_get_cache_files.side_effect = _get_cache_files
-
-    yield tmp_path
-
-    if request.node.rep_call.passed:
-        return
-
-    # Print contents of dir if failed
-    print("Content of dir for", request.node.nodeid)
-    for fil in tmp_path.iterdir():
-        print(fil.relative_to(tmp_path))
-
-    # To show the log.
-    assert False
-
-
-@pytest.fixture()
-def mutagen_mock():
-    """Mock writing tags."""
-    with patch(
-        "homeassistant.components.tts.SpeechManager.write_tags",
-        side_effect=lambda *args: args[1],
-    ):
-        yield
-
-
 @pytest.fixture(autouse=True)
 async def internal_url_mock(hass):
     """Mock internal URL of the instance."""
@@ -730,7 +679,7 @@ async def test_tags_with_wave(hass, demo_provider):
         + "22 56 00 00 88 58 01 00 04 00 10 00 64 61 74 61 00 00 00 00"
     )
 
-    tagged_data = tts.SpeechManager.write_tags(
+    tagged_data = ORIG_WRITE_TAGS(
         "42f18378fd4393d18c8dd11d03fa9563c1e54491_en_-_demo.wav",
         demo_data,
         demo_provider,
diff --git a/tests/components/tts/test_media_source.py b/tests/components/tts/test_media_source.py
new file mode 100644
index 00000000000..3bfd204a228
--- /dev/null
+++ b/tests/components/tts/test_media_source.py
@@ -0,0 +1,99 @@
+"""Tests for TTS media source."""
+from unittest.mock import patch
+
+import pytest
+
+from homeassistant.components import media_source
+from homeassistant.components.media_player.errors import BrowseError
+from homeassistant.setup import async_setup_component
+
+
+@pytest.fixture(autouse=True)
+async def mock_get_tts_audio(hass):
+    """Set up media source."""
+    assert await async_setup_component(hass, "media_source", {})
+    assert await async_setup_component(
+        hass,
+        "tts",
+        {
+            "tts": {
+                "platform": "demo",
+            }
+        },
+    )
+
+    with patch(
+        "homeassistant.components.demo.tts.DemoProvider.get_tts_audio",
+        return_value=("mp3", b""),
+    ) as mock_get_tts:
+        yield mock_get_tts
+
+
+async def test_browsing(hass):
+    """Test browsing TTS media source."""
+    item = await media_source.async_browse_media(hass, "media-source://tts")
+    assert item is not None
+    assert item.title == "Text to Speech"
+    assert len(item.children) == 1
+    assert item.can_play is False
+    assert item.can_expand is True
+
+    item_child = await media_source.async_browse_media(
+        hass, item.children[0].media_content_id
+    )
+    assert item_child is not None
+    assert item_child.title == "Demo"
+    assert item_child.children is None
+    assert item_child.can_play is False
+    assert item_child.can_expand is True
+
+    with pytest.raises(BrowseError):
+        await media_source.async_browse_media(hass, "media-source://tts/non-existing")
+
+
+async def test_resolving(hass, mock_get_tts_audio):
+    """Test resolving."""
+    media = await media_source.async_resolve_media(
+        hass, "media-source://tts/demo?message=Hello%20World"
+    )
+    assert media.url.startswith("/api/tts_proxy/")
+    assert media.mime_type == "audio/mpeg"
+
+    assert len(mock_get_tts_audio.mock_calls) == 1
+    message, language = mock_get_tts_audio.mock_calls[0][1]
+    assert message == "Hello World"
+    assert language == "en"
+    assert mock_get_tts_audio.mock_calls[0][2]["options"] is None
+
+    # Pass language and options
+    mock_get_tts_audio.reset_mock()
+    media = await media_source.async_resolve_media(
+        hass, "media-source://tts/demo?message=Bye%20World&language=de&voice=Paulus"
+    )
+    assert media.url.startswith("/api/tts_proxy/")
+    assert media.mime_type == "audio/mpeg"
+
+    assert len(mock_get_tts_audio.mock_calls) == 1
+    message, language = mock_get_tts_audio.mock_calls[0][1]
+    assert message == "Bye World"
+    assert language == "de"
+    assert mock_get_tts_audio.mock_calls[0][2]["options"] == {"voice": "Paulus"}
+
+
+async def test_resolving_errors(hass):
+    """Test resolving."""
+    # No message added
+    with pytest.raises(media_source.Unresolvable):
+        await media_source.async_resolve_media(hass, "media-source://tts/demo")
+
+    # Non-existing provider
+    with pytest.raises(media_source.Unresolvable):
+        await media_source.async_resolve_media(
+            hass, "media-source://tts/non-existing?message=bla"
+        )
+
+    # Non-existing option
+    with pytest.raises(media_source.Unresolvable):
+        await media_source.async_resolve_media(
+            hass, "media-source://tts/non-existing?message=bla&non_existing_option=bla"
+        )
diff --git a/tests/components/tts/test_notify.py b/tests/components/tts/test_notify.py
index 9989b1d349e..912896dd3e2 100644
--- a/tests/components/tts/test_notify.py
+++ b/tests/components/tts/test_notify.py
@@ -1,6 +1,4 @@
 """The tests for the TTS component."""
-from unittest.mock import patch
-
 import pytest
 import yarl
 
@@ -22,16 +20,6 @@ def relative_url(url):
     return str(yarl.URL(url).relative())
 
 
-@pytest.fixture(autouse=True)
-def mutagen_mock():
-    """Mock writing tags."""
-    with patch(
-        "homeassistant.components.tts.SpeechManager.write_tags",
-        side_effect=lambda *args: args[1],
-    ):
-        yield
-
-
 @pytest.fixture(autouse=True)
 async def internal_url_mock(hass):
     """Mock internal URL of the instance."""
diff --git a/tests/components/voicerss/test_tts.py b/tests/components/voicerss/test_tts.py
index 424bbe5e065..4eab1868057 100644
--- a/tests/components/voicerss/test_tts.py
+++ b/tests/components/voicerss/test_tts.py
@@ -15,7 +15,7 @@ import homeassistant.components.tts as tts
 from homeassistant.setup import async_setup_component
 
 from tests.common import assert_setup_component, async_mock_service
-from tests.components.tts.test_init import mutagen_mock  # noqa: F401
+from tests.components.tts.conftest import mutagen_mock  # noqa: F401
 
 URL = "https://api.voicerss.org/"
 FORM_DATA = {
diff --git a/tests/components/yandextts/test_tts.py b/tests/components/yandextts/test_tts.py
index db4dbce4b8b..495009eecf9 100644
--- a/tests/components/yandextts/test_tts.py
+++ b/tests/components/yandextts/test_tts.py
@@ -14,7 +14,7 @@ import homeassistant.components.tts as tts
 from homeassistant.setup import async_setup_component
 
 from tests.common import assert_setup_component, async_mock_service
-from tests.components.tts.test_init import (  # noqa: F401, pylint: disable=unused-import
+from tests.components.tts.conftest import (  # noqa: F401, pylint: disable=unused-import
     mutagen_mock,
 )