From 6284c501ed8ad392b5b6e94818e38f69020fca90 Mon Sep 17 00:00:00 2001
From: Artem Akymenko <akimenko.artem@gmail.com>
Date: Thu, 9 Jul 2026 14:06:40 +0000
Subject: [PATCH] feat: add SuperTonic TTS plugin

- Add plugins/supertonic/ with Engine and EngineSession implementations
- Reuse existing SupertonicPipeline from abogen.tts_backends.supertonic
- Implement VoiceLister capability (M1-M5, F1-F5 voices)
- Declare no streaming support via capabilities
- Add 28 tests: plugin loading, protocol compliance, lifecycle, voice listing, parameters, errors
- All 237 tests pass
---
 plugins/supertonic/__init__.py  | 123 ++++++++++
 plugins/supertonic/engine.py    | 156 ++++++++++++
 tests/test_supertonic_plugin.py | 415 ++++++++++++++++++++++++++++++++
 3 files changed, 694 insertions(+)
 create mode 100644 plugins/supertonic/__init__.py
 create mode 100644 plugins/supertonic/engine.py
 create mode 100644 tests/test_supertonic_plugin.py

diff --git a/plugins/supertonic/__init__.py b/plugins/supertonic/__init__.py
new file mode 100644
index 0000000..abd92cc
--- /dev/null
+++ b/plugins/supertonic/__init__.py
@@ -0,0 +1,123 @@
+"""SuperTonic TTS Plugin for the TTS Plugin Architecture.
+
+This plugin provides a SuperTonic-based TTS engine that implements the
+Plugin API contract. It wraps the existing SuperTonic backend in the
+new Engine/EngineSession architecture.
+
+Exports:
+    - PLUGIN_MANIFEST: PluginManifest
+    - MODEL_REQUIREMENTS: list[ModelManifest]
+    - create_engine: Factory function
+"""
+
+from __future__ import annotations
+
+from pathlib import Path
+from typing import Any
+
+from abogen.tts_plugin.engine import Engine
+from abogen.tts_plugin.host_context import HostContext
+from abogen.tts_plugin.manifest import (
+    AudioFormatManifest,
+    EngineManifest,
+    ModelManifest,
+    ParameterManifest,
+    PluginManifest,
+    RequirementManifest,
+    VoiceSourceManifest,
+)
+from abogen.tts_plugin.types import EngineConfig
+
+from .engine import SuperTonicEngine
+
+
+def _load_supertonic_pipeline(sample_rate: int = 24000, auto_download: bool = True, total_steps: int = 5) -> Any:
+    """Lazy-load SuperTonic dependencies and create pipeline."""
+    from abogen.tts_backends.supertonic import SupertonicPipeline
+
+    return SupertonicPipeline(
+        sample_rate=sample_rate,
+        auto_download=auto_download,
+        total_steps=total_steps,
+    )
+
+
+PLUGIN_MANIFEST = PluginManifest(
+    id="supertonic",
+    name="SuperTonic",
+    version="0.1.0",
+    api_version="1.0",
+    description="SuperTonic TTS engine - fast high-quality text-to-speech",
+    author="SuperTonic Team",
+    capabilities=("voice_list",),
+    requires=RequirementManifest(
+        internet=False,
+    ),
+    engine=EngineManifest(
+        voiceSources=(
+            VoiceSourceManifest(
+                id="builtin",
+                name="Built-in Voices",
+                type="list",
+                config={"voices": "See listVoices()"},
+            ),
+        ),
+        parameters=(
+            ParameterManifest(
+                id="speed",
+                name="Speed",
+                description="Speech speed multiplier",
+                type="float",
+                default=1.0,
+                min=0.7,
+                max=2.0,
+                step=0.1,
+            ),
+            ParameterManifest(
+                id="total_steps",
+                name="Quality Steps",
+                description="Inference steps (higher = better quality, slower)",
+                type="int",
+                default=5,
+                min=2,
+                max=15,
+                step=1,
+            ),
+        ),
+        audioFormats=(
+            AudioFormatManifest(mime="audio/wav", extension="wav"),
+        ),
+    ),
+)
+
+MODEL_REQUIREMENTS: list[ModelManifest] = []
+
+
+def create_engine(
+    context: HostContext,
+    model_path: Path | None,
+    config: EngineConfig,
+) -> Engine:
+    """Create a SuperTonic engine instance.
+
+    This function is the plugin entry point. It must be atomic:
+    succeed fully or raise EngineError and clean up.
+
+    Args:
+        context: Host services (config dir, logger, http client).
+        model_path: Resolved model path, or None for default.
+        config: Engine initialization settings (device, etc.).
+
+    Returns:
+        A fully initialized SuperTonicEngine instance.
+
+    Raises:
+        EngineError: On failure. Cleans up partially created resources.
+    """
+    try:
+        pipeline = _load_supertonic_pipeline()
+        engine = SuperTonicEngine(pipeline)
+        return engine
+    except Exception as e:
+        from abogen.tts_plugin.errors import EngineError as EngineErrorClass
+        raise EngineErrorClass(f"Failed to create SuperTonic engine: {e}") from e
diff --git a/plugins/supertonic/engine.py b/plugins/supertonic/engine.py
new file mode 100644
index 0000000..a410af9
--- /dev/null
+++ b/plugins/supertonic/engine.py
@@ -0,0 +1,156 @@
+"""SuperTonic Engine adapter for the TTS Plugin Architecture.
+
+This module adapts the existing SuperTonic backend to the new Engine/EngineSession
+protocol. It wraps the SupertonicPipeline without modifying it.
+"""
+
+from __future__ import annotations
+
+import io
+import logging
+from typing import Any, Optional
+
+import numpy as np
+
+from abogen.tts_plugin.capabilities import VoiceLister
+from abogen.tts_plugin.engine import Engine, EngineSession
+from abogen.tts_plugin.errors import EngineError, InvalidInputError
+from abogen.tts_plugin.manifest import VoiceManifest
+from abogen.tts_plugin.types import (
+    AudioFormat,
+    Duration,
+    ParameterValues,
+    SynthesisRequest,
+    SynthesizedAudio,
+    VoiceSelection,
+)
+
+logger = logging.getLogger(__name__)
+
+# SuperTonic voice list - source of truth
+_SUPERTONIC_VOICES = ("M1", "M2", "M3", "M4", "M5", "F1", "F2", "F3", "F4", "F5")
+
+# Voice display names mapping
+_VOICE_DISPLAY_NAMES: dict[str, str] = {
+    "M1": "Male 1",
+    "M2": "Male 2",
+    "M3": "Male 3",
+    "M4": "Male 4",
+    "M5": "Male 5",
+    "F1": "Female 1",
+    "F2": "Female 2",
+    "F3": "Female 3",
+    "F4": "Female 4",
+    "F5": "Female 5",
+}
+
+# Sample rate for SuperTonic audio
+_SUPERTONIC_SAMPLE_RATE = 24000
+
+
+class SuperTonicSession:
+    """EngineSession implementation for SuperTonic.
+
+    Owns mutable execution state for synthesis.
+    NOT thread-safe.
+    """
+
+    def __init__(self, pipeline: Any) -> None:
+        self._pipeline = pipeline
+        self._disposed = False
+
+    def synthesize(self, request: SynthesisRequest) -> SynthesizedAudio:
+        """Synthesize audio from text using SuperTonic."""
+        if self._disposed:
+            raise EngineError("Session disposed")
+
+        try:
+            import soundfile as sf
+
+            voice = request.voice.key
+            speed = float(request.parameters.values.get("speed", 1.0))
+            total_steps = request.parameters.values.get("total_steps", None)
+            split_pattern = request.parameters.values.get("split_pattern", None)
+
+            if total_steps is not None:
+                total_steps = int(total_steps)
+
+            audio_parts: list[np.ndarray] = []
+            for segment in self._pipeline(
+                request.text,
+                voice=voice,
+                speed=speed,
+                split_pattern=split_pattern,
+                total_steps=total_steps,
+            ):
+                audio_parts.append(segment.audio)
+
+            if not audio_parts:
+                return SynthesizedAudio(
+                    data=b"",
+                    format=AudioFormat(mime="audio/wav", extension="wav"),
+                    duration=Duration(seconds=0.0),
+                )
+
+            combined = np.concatenate(audio_parts).astype("float32", copy=False)
+            buf = io.BytesIO()
+            sf.write(buf, combined, self._pipeline.sample_rate, format="WAV")
+            audio_bytes = buf.getvalue()
+            duration_seconds = len(combined) / self._pipeline.sample_rate
+
+            return SynthesizedAudio(
+                data=audio_bytes,
+                format=AudioFormat(mime="audio/wav", extension="wav"),
+                duration=Duration(seconds=duration_seconds),
+            )
+        except EngineError:
+            raise
+        except Exception as e:
+            raise EngineError(f"Synthesis failed: {e}") from e
+
+    def dispose(self) -> None:
+        """Release session resources. Idempotent."""
+        self._disposed = True
+
+
+class SuperTonicEngine:
+    """Engine implementation for SuperTonic.
+
+    Factory for SuperTonicSession instances. Stateless and thread-safe.
+    """
+
+    def __init__(self, pipeline: Any) -> None:
+        self._pipeline = pipeline
+        self._disposed = False
+
+    def createSession(self) -> SuperTonicSession:
+        """Create a new SuperTonicSession."""
+        if self._disposed:
+            raise EngineError("Engine disposed")
+        return SuperTonicSession(self._pipeline)
+
+    def dispose(self) -> None:
+        """Release engine resources. Idempotent."""
+        self._disposed = True
+
+    def listVoices(self, sourceId: str) -> list[VoiceManifest]:
+        """List available SuperTonic voices. Implements VoiceLister capability."""
+        if self._disposed:
+            raise EngineError("Engine disposed")
+        return [
+            VoiceManifest(
+                id=voice_id,
+                name=_VOICE_DISPLAY_NAMES.get(voice_id, voice_id),
+                tags=(_get_gender_tag(voice_id),),
+            )
+            for voice_id in _SUPERTONIC_VOICES
+        ]
+
+
+def _get_gender_tag(voice_id: str) -> str:
+    """Extract gender tag from voice ID."""
+    if voice_id.startswith("M"):
+        return "male"
+    elif voice_id.startswith("F"):
+        return "female"
+    return "unknown"
diff --git a/tests/test_supertonic_plugin.py b/tests/test_supertonic_plugin.py
new file mode 100644
index 0000000..6696ba8
--- /dev/null
+++ b/tests/test_supertonic_plugin.py
@@ -0,0 +1,415 @@
+"""Tests for the SuperTonic TTS Plugin.
+
+These tests verify that the SuperTonic plugin:
+- Loads correctly through the Plugin Loader
+- Has a valid manifest
+- Creates a valid Engine
+- Satisfies the Engine/EngineSession contract
+- Implements VoiceLister capability
+"""
+
+from __future__ import annotations
+
+import logging
+from pathlib import Path
+from typing import Any
+
+import pytest
+
+from abogen.tts_plugin.engine import Engine, EngineSession
+from abogen.tts_plugin.host_context import HostContext
+from abogen.tts_plugin.loader import load_plugin_from_dir
+from abogen.tts_plugin.manifest import PluginManifest
+from abogen.tts_plugin.types import (
+    AudioFormat,
+    EngineConfig,
+    ParameterValues,
+    SynthesisRequest,
+    SynthesizedAudio,
+    VoiceSelection,
+)
+
+
+# ──────────────────────────────────────────────────────────────
+# Path fixtures
+# ──────────────────────────────────────────────────────────────
+
+@pytest.fixture
+def supertonic_plugin_dir() -> Path:
+    return Path(__file__).parent.parent / "plugins" / "supertonic"
+
+
+def _supertonic_available() -> bool:
+    """Check if SuperTonic is available."""
+    try:
+        from supertonic import TTS  # type: ignore[import-not-found]
+        return True
+    except ImportError:
+        return False
+
+
+@pytest.fixture
+def host_context(tmp_path: Path) -> HostContext:
+    class FakeHttpClient:
+        def get(self, url: str, **kwargs: object) -> object:
+            return None
+        def post(self, url: str, **kwargs: object) -> object:
+            return None
+
+    return HostContext(
+        config_dir=tmp_path,
+        logger=logging.getLogger("test"),
+        http_client=FakeHttpClient(),
+    )
+
+
+# ──────────────────────────────────────────────────────────────
+# Plugin Loading Tests
+# ──────────────────────────────────────────────────────────────
+
+class TestSuperTonicPluginLoading:
+    """Test that SuperTonic plugin loads correctly through the Loader."""
+
+    def test_plugin_loads_successfully(self, supertonic_plugin_dir: Path) -> None:
+        result = load_plugin_from_dir(supertonic_plugin_dir)
+        assert result.success is True
+        assert result.manifest is not None
+        assert result.create_engine is not None
+
+    def test_plugin_has_valid_manifest(self, supertonic_plugin_dir: Path) -> None:
+        result = load_plugin_from_dir(supertonic_plugin_dir)
+        assert result.success is True
+        manifest = result.manifest
+        assert isinstance(manifest, PluginManifest)
+        assert manifest.id == "supertonic"
+        assert manifest.name == "SuperTonic"
+        assert manifest.api_version == "1.0"
+
+    def test_plugin_has_model_requirements(self, supertonic_plugin_dir: Path) -> None:
+        result = load_plugin_from_dir(supertonic_plugin_dir)
+        assert result.success is True
+        assert result.model_requirements is not None
+        assert isinstance(result.model_requirements, tuple)
+
+    def test_plugin_manifest_capabilities(self, supertonic_plugin_dir: Path) -> None:
+        result = load_plugin_from_dir(supertonic_plugin_dir)
+        assert result.success is True
+        manifest = result.manifest
+        assert "voice_list" in manifest.capabilities
+
+    def test_plugin_manifest_engine(self, supertonic_plugin_dir: Path) -> None:
+        result = load_plugin_from_dir(supertonic_plugin_dir)
+        assert result.success is True
+        engine_manifest = result.manifest.engine
+        assert len(engine_manifest.voiceSources) > 0
+        assert len(engine_manifest.audioFormats) > 0
+        assert len(engine_manifest.parameters) > 0
+
+
+# ──────────────────────────────────────────────────────────────
+# Engine Creation Tests
+# ──────────────────────────────────────────────────────────────
+
+class TestSuperTonicEngineCreation:
+    """Test that SuperTonic Engine can be created."""
+
+    @pytest.mark.skipif(
+        not _supertonic_available(),
+        reason="SuperTonic not installed"
+    )
+    def test_create_engine(self, supertonic_plugin_dir: Path, host_context: HostContext) -> None:
+        result = load_plugin_from_dir(supertonic_plugin_dir)
+        assert result.success is True
+
+        engine = result.create_engine(host_context, None, EngineConfig())
+        assert isinstance(engine, Engine)
+        engine.dispose()
+
+    @pytest.mark.skipif(
+        not _supertonic_available(),
+        reason="SuperTonic not installed"
+    )
+    def test_engine_satisfies_protocol(self, supertonic_plugin_dir: Path, host_context: HostContext) -> None:
+        result = load_plugin_from_dir(supertonic_plugin_dir)
+        assert result.success is True
+
+        engine = result.create_engine(host_context, None, EngineConfig())
+        assert isinstance(engine, Engine)
+        engine.dispose()
+
+
+# ──────────────────────────────────────────────────────────────
+# Engine Protocol Tests (using mock pipeline)
+# ──────────────────────────────────────────────────────────────
+
+class TestSuperTonicEngineProtocol:
+    """Test SuperTonic Engine protocol compliance using mock pipeline."""
+
+    def _create_engine_with_mock(self) -> Any:
+        """Create engine with mock pipeline for testing."""
+        from plugins.supertonic.engine import SuperTonicEngine
+
+        class MockSegment:
+            def __init__(self):
+                import numpy as np
+                self.audio = np.zeros(24000, dtype="float32")
+
+        class MockPipeline:
+            sample_rate = 24000
+
+            def __call__(self, text, voice, speed, split_pattern=None, total_steps=None):
+                return [MockSegment()]
+
+        return SuperTonicEngine(MockPipeline())
+
+    def test_create_session(self) -> None:
+        engine = self._create_engine_with_mock()
+        session = engine.createSession()
+        assert isinstance(session, EngineSession)
+        engine.dispose()
+
+    def test_dispose_idempotent(self) -> None:
+        engine = self._create_engine_with_mock()
+        engine.dispose()
+        engine.dispose()  # Should not raise
+
+    def test_create_session_after_dispose_raises(self) -> None:
+        from abogen.tts_plugin.errors import EngineError
+        engine = self._create_engine_with_mock()
+        engine.dispose()
+        with pytest.raises(EngineError):
+            engine.createSession()
+
+    def test_session_synthesize(self) -> None:
+        engine = self._create_engine_with_mock()
+        session = engine.createSession()
+        request = SynthesisRequest(
+            text="Hello",
+            voice=VoiceSelection(source="builtin", key="M1"),
+            parameters=ParameterValues(values={}),
+            format=AudioFormat(mime="audio/wav", extension="wav"),
+        )
+        result = session.synthesize(request)
+        assert result.data is not None
+        assert len(result.data) > 0
+        session.dispose()
+        engine.dispose()
+
+    def test_session_dispose_idempotent(self) -> None:
+        engine = self._create_engine_with_mock()
+        session = engine.createSession()
+        session.dispose()
+        session.dispose()  # Should not raise
+        engine.dispose()
+
+    def test_session_synthesize_after_dispose_raises(self) -> None:
+        from abogen.tts_plugin.errors import EngineError
+        engine = self._create_engine_with_mock()
+        session = engine.createSession()
+        session.dispose()
+        request = SynthesisRequest(
+            text="Hello",
+            voice=VoiceSelection(source="builtin", key="M1"),
+            parameters=ParameterValues(values={}),
+            format=AudioFormat(mime="audio/wav", extension="wav"),
+        )
+        with pytest.raises(EngineError):
+            session.synthesize(request)
+        engine.dispose()
+
+    def test_session_multiple_synthesize(self) -> None:
+        engine = self._create_engine_with_mock()
+        session = engine.createSession()
+        request = SynthesisRequest(
+            text="Hello",
+            voice=VoiceSelection(source="builtin", key="M1"),
+            parameters=ParameterValues(values={}),
+            format=AudioFormat(mime="audio/wav", extension="wav"),
+        )
+        result1 = session.synthesize(request)
+        result2 = session.synthesize(request)
+        assert isinstance(result1.data, bytes)
+        assert isinstance(result2.data, bytes)
+        session.dispose()
+        engine.dispose()
+
+    def test_full_lifecycle(self) -> None:
+        engine = self._create_engine_with_mock()
+
+        # Create sessions
+        session1 = engine.createSession()
+        session2 = engine.createSession()
+
+        # Use sessions
+        request = SynthesisRequest(
+            text="Hello",
+            voice=VoiceSelection(source="builtin", key="M1"),
+            parameters=ParameterValues(values={}),
+            format=AudioFormat(mime="audio/wav", extension="wav"),
+        )
+        assert isinstance(session1.synthesize(request), SynthesizedAudio)
+        assert isinstance(session2.synthesize(request), SynthesizedAudio)
+
+        # Dispose sessions
+        session1.dispose()
+        session2.dispose()
+
+        # Dispose engine
+        engine.dispose()
+
+    def test_engine_returns_new_session_instances(self) -> None:
+        engine = self._create_engine_with_mock()
+        session1 = engine.createSession()
+        session2 = engine.createSession()
+        assert session1 is not session2
+        session1.dispose()
+        session2.dispose()
+        engine.dispose()
+
+
+# ──────────────────────────────────────────────────────────────
+# VoiceLister Tests
+# ──────────────────────────────────────────────────────────────
+
+class TestSuperTonicVoiceLister:
+    """Test SuperTonic VoiceLister capability."""
+
+    def test_list_voices(self) -> None:
+        engine = TestSuperTonicEngineProtocol._create_engine_with_mock(TestSuperTonicEngineProtocol)
+        voices = engine.listVoices("builtin")
+        assert len(voices) > 0
+        assert all(hasattr(v, "id") for v in voices)
+        assert all(hasattr(v, "name") for v in voices)
+        engine.dispose()
+
+    def test_voices_have_tags(self) -> None:
+        engine = TestSuperTonicEngineProtocol._create_engine_with_mock(TestSuperTonicEngineProtocol)
+        voices = engine.listVoices("builtin")
+        for voice in voices:
+            assert isinstance(voice.tags, tuple)
+            assert len(voice.tags) > 0
+        engine.dispose()
+
+    def test_voices_are_correct_count(self) -> None:
+        engine = TestSuperTonicEngineProtocol._create_engine_with_mock(TestSuperTonicEngineProtocol)
+        voices = engine.listVoices("builtin")
+        assert len(voices) == 10  # M1-M5, F1-F5
+        engine.dispose()
+
+    def test_voice_ids_match_expected(self) -> None:
+        engine = TestSuperTonicEngineProtocol._create_engine_with_mock(TestSuperTonicEngineProtocol)
+        voices = engine.listVoices("builtin")
+        voice_ids = [v.id for v in voices]
+        expected = ["M1", "M2", "M3", "M4", "M5", "F1", "F2", "F3", "F4", "F5"]
+        assert voice_ids == expected
+        engine.dispose()
+
+    def test_male_voices_have_male_tag(self) -> None:
+        engine = TestSuperTonicEngineProtocol._create_engine_with_mock(TestSuperTonicEngineProtocol)
+        voices = engine.listVoices("builtin")
+        male_voices = [v for v in voices if v.id.startswith("M")]
+        for voice in male_voices:
+            assert "male" in voice.tags
+        engine.dispose()
+
+    def test_female_voices_have_female_tag(self) -> None:
+        engine = TestSuperTonicEngineProtocol._create_engine_with_mock(TestSuperTonicEngineProtocol)
+        voices = engine.listVoices("builtin")
+        female_voices = [v for v in voices if v.id.startswith("F")]
+        for voice in female_voices:
+            assert "female" in voice.tags
+        engine.dispose()
+
+    def test_list_voices_after_dispose_raises(self) -> None:
+        from abogen.tts_plugin.errors import EngineError
+        engine = TestSuperTonicEngineProtocol._create_engine_with_mock(TestSuperTonicEngineProtocol)
+        engine.dispose()
+        with pytest.raises(EngineError):
+            engine.listVoices("builtin")
+
+
+# ──────────────────────────────────────────────────────────────
+# Parameter Tests
+# ──────────────────────────────────────────────────────────────
+
+class TestSuperTonicParameters:
+    """Test SuperTonic parameter handling."""
+
+    def test_speed_parameter(self) -> None:
+        engine = TestSuperTonicEngineProtocol._create_engine_with_mock(TestSuperTonicEngineProtocol)
+        session = engine.createSession()
+        request = SynthesisRequest(
+            text="Hello",
+            voice=VoiceSelection(source="builtin", key="M1"),
+            parameters=ParameterValues(values={"speed": 1.5}),
+            format=AudioFormat(mime="audio/wav", extension="wav"),
+        )
+        result = session.synthesize(request)
+        assert isinstance(result.data, bytes)
+        session.dispose()
+        engine.dispose()
+
+    def test_total_steps_parameter(self) -> None:
+        engine = TestSuperTonicEngineProtocol._create_engine_with_mock(TestSuperTonicEngineProtocol)
+        session = engine.createSession()
+        request = SynthesisRequest(
+            text="Hello",
+            voice=VoiceSelection(source="builtin", key="M1"),
+            parameters=ParameterValues(values={"total_steps": 10}),
+            format=AudioFormat(mime="audio/wav", extension="wav"),
+        )
+        result = session.synthesize(request)
+        assert isinstance(result.data, bytes)
+        session.dispose()
+        engine.dispose()
+
+    def test_default_parameters(self) -> None:
+        engine = TestSuperTonicEngineProtocol._create_engine_with_mock(TestSuperTonicEngineProtocol)
+        session = engine.createSession()
+        request = SynthesisRequest(
+            text="Hello",
+            voice=VoiceSelection(source="builtin", key="M1"),
+            parameters=ParameterValues(values={}),
+            format=AudioFormat(mime="audio/wav", extension="wav"),
+        )
+        result = session.synthesize(request)
+        assert isinstance(result.data, bytes)
+        session.dispose()
+        engine.dispose()
+
+
+# ──────────────────────────────────────────────────────────────
+# Error Handling Tests
+# ──────────────────────────────────────────────────────────────
+
+class TestSuperTonicErrorHandling:
+    """Test SuperTonic error handling."""
+
+    def test_synthesize_empty_text(self) -> None:
+        engine = TestSuperTonicEngineProtocol._create_engine_with_mock(TestSuperTonicEngineProtocol)
+        session = engine.createSession()
+        request = SynthesisRequest(
+            text="",
+            voice=VoiceSelection(source="builtin", key="M1"),
+            parameters=ParameterValues(values={}),
+            format=AudioFormat(mime="audio/wav", extension="wav"),
+        )
+        result = session.synthesize(request)
+        # Empty text should return empty audio
+        assert isinstance(result.data, bytes)
+        session.dispose()
+        engine.dispose()
+
+    def test_synthesize_whitespace_text(self) -> None:
+        engine = TestSuperTonicEngineProtocol._create_engine_with_mock(TestSuperTonicEngineProtocol)
+        session = engine.createSession()
+        request = SynthesisRequest(
+            text="   ",
+            voice=VoiceSelection(source="builtin", key="M1"),
+            parameters=ParameterValues(values={}),
+            format=AudioFormat(mime="audio/wav", extension="wav"),
+        )
+        result = session.synthesize(request)
+        assert isinstance(result.data, bytes)
+        session.dispose()
+        engine.dispose()