agent-cli 0.79.0__tar.gz → 0.80.0__tar.gz

{agent_cli-0.79.0 → agent_cli-0.80.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agent-cli
-Version: 0.79.0
+Version: 0.80.0
 Summary: A suite of AI-powered command-line tools for text correction, audio transcription, and voice assistance.
 Project-URL: Homepage, https://github.com/basnijholt/agent-cli
 Author-email: Bas Nijholt <bas@nijho.lt>

{agent_cli-0.79.0 → agent_cli-0.80.0}/agent_cli/dev/cli.py

@@ -1469,6 +1469,8 @@ def _clean_merged_worktrees(
     repo_root: Path,
     dry_run: bool,
     yes: bool,
+    *,
+    force: bool = False,
 ) -> None:
     """Remove worktrees with merged PRs (requires gh CLI)."""
     _info("Checking for worktrees with merged PRs...")
@@ -1509,7 +1511,7 @@ def _clean_merged_worktrees(
         for wt, _pr_url in to_remove:
             success, error = worktree.remove_worktree(
                 wt.path,
-                force=False,
+                force=force,
                 delete_branch=True,
                 repo_path=repo_root,
             )
@@ -1523,6 +1525,8 @@ def _clean_no_commits_worktrees(
     repo_root: Path,
     dry_run: bool,
     yes: bool,
+    *,
+    force: bool = False,
 ) -> None:
     """Remove worktrees with no commits ahead of the default branch."""
     _info("Checking for worktrees with no commits...")
@@ -1546,7 +1550,7 @@ def _clean_no_commits_worktrees(
         for wt in to_remove:
             success, error = worktree.remove_worktree(
                 wt.path,
-                force=False,
+                force=force,
                 delete_branch=True,
                 repo_path=repo_root,
             )
@@ -1584,6 +1588,14 @@ def clean(
         bool,
         typer.Option("--yes", "-y", help="Skip confirmation prompts"),
     ] = False,
+    force: Annotated[
+        bool,
+        typer.Option(
+            "--force",
+            "-f",
+            help="Force removal of worktrees with modified or untracked files",
+        ),
+    ] = False,
 ) -> None:
     """Clean up stale worktrees and empty directories.
 
@@ -1601,6 +1613,7 @@ def clean(
     - `dev clean` — Basic cleanup
     - `dev clean --merged` — Remove worktrees with merged PRs
     - `dev clean --merged --dry-run` — Preview what would be removed
+    - `dev clean --no-commits --force` — Force remove abandoned worktrees with local changes
     """
     repo_root = _ensure_git_repo()
 
@@ -1635,11 +1648,11 @@ def clean(
 
     # --merged mode: remove worktrees with merged PRs
     if merged:
-        _clean_merged_worktrees(repo_root, dry_run, yes)
+        _clean_merged_worktrees(repo_root, dry_run, yes, force=force)
 
     # --no-commits mode: remove worktrees with no commits ahead of default branch
     if no_commits:
-        _clean_no_commits_worktrees(repo_root, dry_run, yes)
+        _clean_no_commits_worktrees(repo_root, dry_run, yes, force=force)
 
 
 @app.command("doctor")

{agent_cli-0.79.0 → agent_cli-0.80.0}/agent_cli/scripts/setup-macos.sh

@@ -47,15 +47,9 @@ echo "⏳ This may take a few minutes depending on your internet connection..."
 sleep 2  # Give Ollama service time to start
 ollama pull gemma3:4b
 
-# Install wyoming-mlx-whisper as a launchd service (Apple Silicon only)
-if [ "$(uname -m)" = "arm64" ]; then
-    echo "🎤 Installing wyoming-mlx-whisper as a background service..."
-    echo "   This will run speech-to-text on Apple Silicon using MLX"
-    curl -fsSL https://raw.githubusercontent.com/basnijholt/wyoming-mlx-whisper/main/scripts/install_service.sh | bash
-    echo "✅ wyoming-mlx-whisper installed as launchd service"
-else
-    echo "ℹ️ Skipping wyoming-mlx-whisper service (Intel Mac - use Linux-style setup)"
-fi
+# Install whisper and tts as launchd daemons
+echo "🎤 Installing whisper and tts-kokoro as background daemons..."
+agent-cli daemon install whisper tts-kokoro -y
 
 echo ""
 echo "✅ Setup complete! You can now run the services:"
@@ -65,12 +59,8 @@ echo "  ./start-all-services.sh"
 echo ""
 echo "Option 2 - Run services individually:"
 echo "  1. Ollama: running as brew service (brew services start ollama)"
-if [ "$(uname -m)" = "arm64" ]; then
-    echo "  2. Whisper: running as launchd service (wyoming-mlx-whisper)"
-else
-    echo "  2. Whisper: agent-cli server whisper"
-fi
-echo "  3. Piper: agent-cli server tts --backend piper"
+echo "  2. Whisper: running as launchd daemon (agent-cli daemon status whisper)"
+echo "  3. TTS: running as launchd daemon (agent-cli daemon status tts-kokoro)"
 echo "  4. OpenWakeWord: ./run-openwakeword.sh"
 echo ""
 echo "🎉 agent-cli has been installed and is ready to use!"

{agent_cli-0.79.0 → agent_cli-0.80.0}/agent_cli/server/tts/backends/piper.py

@@ -1,4 +1,4 @@
-"""Piper TTS backend using piper-tts library."""
+"""Piper TTS backend using piper-tts library with subprocess isolation."""
 
 from __future__ import annotations
 
@@ -7,10 +7,14 @@ import io
 import logging
 import time
 import wave
+from concurrent.futures import ProcessPoolExecutor
+from dataclasses import dataclass
+from multiprocessing import get_context
 from pathlib import Path
-from typing import TYPE_CHECKING, Any, NoReturn
+from typing import Any, NoReturn
 
 from agent_cli import constants
+from agent_cli.core.process import set_process_title
 from agent_cli.server.tts.backends.base import (
     BackendConfig,
     InvalidTextError,
@@ -18,29 +22,45 @@ from agent_cli.server.tts.backends.base import (
     get_backend_cache_dir,
 )
 
-if TYPE_CHECKING:
-    from piper import PiperVoice
-
 logger = logging.getLogger(__name__)
 
 
-def _load_model_sync(
+# --- Subprocess state (only used within subprocess worker) ---
+# This state persists across function calls within the subprocess because:
+# 1. Model loading is expensive and must be reused across synthesis calls
+# 2. The subprocess is long-lived (ProcessPoolExecutor reuses workers)
+
+
+@dataclass
+class _SubprocessState:
+    """Container for subprocess-local state. Not shared with main process."""
+
+    voice: Any = None
+    sample_rate: int = constants.PIPER_DEFAULT_SAMPLE_RATE
+
+
+_state = _SubprocessState()
+
+
+def _load_model_in_subprocess(
     model_name: str,
     cache_dir: str | None,
-) -> tuple[Any, int]:
-    """Load Piper model synchronously (for use in process pool).
+) -> int:
+    """Load Piper model in subprocess. Returns sample rate.
 
     Args:
         model_name: Model name (e.g., 'en_US-lessac-medium') or path to .onnx file.
         cache_dir: Optional cache directory for downloaded models.
 
     Returns:
-        Tuple of (PiperVoice, sample_rate).
+        Sample rate of loaded model.
 
     """
     from piper import PiperVoice  # noqa: PLC0415
     from piper.download_voices import download_voice  # noqa: PLC0415
 
+    set_process_title("tts-piper")
+
     # Use default cache dir if not specified
     download_dir = Path(cache_dir) if cache_dir else get_backend_cache_dir("piper")
     download_dir.mkdir(parents=True, exist_ok=True)
@@ -50,7 +70,9 @@ def _load_model_sync(
     if model_path.exists() and model_path.suffix == ".onnx":
         # Direct path to model file
         voice = PiperVoice.load(str(model_path), use_cuda=False)
-        return voice, voice.config.sample_rate
+        _state.voice = voice
+        _state.sample_rate = voice.config.sample_rate
+        return _state.sample_rate
 
     # Otherwise, treat as a voice name and download if needed
     voice_code = model_name.strip()
@@ -60,24 +82,22 @@ def _load_model_sync(
         logger.info("Downloading Piper voice: %s", voice_code)
         download_voice(voice_code, download_dir)
 
-    # Load the voice
+    # Load the voice and store in subprocess state
     voice = PiperVoice.load(str(expected_model_path), use_cuda=False)
+    _state.voice = voice
+    _state.sample_rate = voice.config.sample_rate
 
-    return voice, voice.config.sample_rate
+    return _state.sample_rate
 
 
-def _synthesize_sync(
-    voice: PiperVoice,
+def _synthesize_in_subprocess(
     text: str,
-    sample_rate: int,
     length_scale: float,
 ) -> tuple[bytes, float]:
-    """Synthesize text to audio synchronously.
+    """Synthesize text to audio in subprocess. Uses model from _state.
 
     Args:
-        voice: Loaded PiperVoice instance.
         text: Text to synthesize.
-        sample_rate: Sample rate from model config.
         length_scale: Length scale (inverse of speed).
 
     Returns:
@@ -86,6 +106,10 @@ def _synthesize_sync(
     """
     from piper import SynthesisConfig  # noqa: PLC0415
 
+    if _state.voice is None:
+        msg = "Model not loaded in subprocess. Call _load_model_in_subprocess first."
+        raise RuntimeError(msg)
+
     # Create synthesis config with speed adjustment
     syn_config = SynthesisConfig(length_scale=length_scale)
 
@@ -94,91 +118,83 @@ def _synthesize_sync(
     with wave.open(buffer, "wb") as wav_file:
         wav_file.setnchannels(1)
         wav_file.setsampwidth(2)  # 16-bit
-        wav_file.setframerate(sample_rate)
+        wav_file.setframerate(_state.sample_rate)
 
         # Synthesize and write audio chunks
-        for audio_chunk in voice.synthesize(text, syn_config):
+        for audio_chunk in _state.voice.synthesize(text, syn_config):
             wav_file.writeframes(audio_chunk.audio_int16_bytes)
 
     audio_data = buffer.getvalue()
 
     # Calculate duration: PCM data size / (sample_rate * channels * bytes_per_sample)
     data_size = len(audio_data) - constants.WAV_HEADER_SIZE
-    duration = data_size / (sample_rate * 1 * 2)
+    duration = data_size / (_state.sample_rate * 1 * 2)
 
     return audio_data, duration
 
 
 class PiperBackend:
-    """Piper TTS backend using ONNX-based synthesis.
+    """Piper TTS backend with subprocess isolation.
 
-    This backend uses the piper-tts library for fast, CPU-friendly TTS.
-    Models are downloaded from HuggingFace on first use.
+    Uses subprocess isolation: when unloaded, the subprocess terminates
+    and all memory is freed by the OS. This ensures clean memory management.
     """
 
     def __init__(self, config: BackendConfig) -> None:
-        """Initialize the Piper backend.
-
-        Args:
-            config: Backend configuration.
-
-        """
+        """Initialize the Piper backend."""
         self._config = config
-        self._voice: PiperVoice | None = None
+        self._executor: ProcessPoolExecutor | None = None
         self._sample_rate: int = constants.PIPER_DEFAULT_SAMPLE_RATE  # Updated on load
         self._device: str | None = None
 
     @property
     def is_loaded(self) -> bool:
-        """Check if the model is currently loaded."""
-        return self._voice is not None
+        """Check if the model is loaded."""
+        return self._executor is not None
 
     @property
     def device(self) -> str | None:
-        """Get the device the model is loaded on, or None if not loaded."""
+        """Get the device the model is on."""
         return self._device
 
     async def load(self) -> float:
-        """Load the model into memory.
-
-        Returns:
-            Load duration in seconds.
-
-        """
-        if self._voice is not None:
+        """Start subprocess and load model."""
+        if self._executor is not None:
             return 0.0
 
         start_time = time.time()
 
-        # Load synchronously since Piper is fast and CPU-only
+        # Subprocess isolation: spawn context for clean state
+        ctx = get_context("spawn")
+        self._executor = ProcessPoolExecutor(max_workers=1, mp_context=ctx)
+
         loop = asyncio.get_running_loop()
-        voice, sample_rate = await loop.run_in_executor(
-            None,
-            _load_model_sync,
+        self._sample_rate = await loop.run_in_executor(
+            self._executor,
+            _load_model_in_subprocess,
             self._config.model_name,
             str(self._config.cache_dir) if self._config.cache_dir else None,
         )
 
-        self._voice = voice
-        self._sample_rate = sample_rate
         self._device = "cpu"  # Piper is CPU-only
 
         load_duration = time.time() - start_time
         logger.info(
-            "Loaded Piper model %s in %.2fs (sample_rate=%d)",
+            "Loaded Piper model %s on %s in %.2fs",
             self._config.model_name,
+            self._device,
             load_duration,
-            self._sample_rate,
         )
-
         return load_duration
 
     async def unload(self) -> None:
-        """Unload the model and free memory."""
-        if self._voice is not None:
-            logger.info("Unloading Piper model %s", self._config.model_name)
-            self._voice = None
-            self._device = None
+        """Shutdown subprocess, releasing all memory."""
+        if self._executor is None:
+            return
+        self._executor.shutdown(wait=False, cancel_futures=True)
+        self._executor = None
+        self._device = None
+        logger.info("Piper model %s unloaded (subprocess terminated)", self._config.model_name)
 
     async def synthesize(
         self,
@@ -187,23 +203,9 @@ class PiperBackend:
         voice: str | None = None,  # noqa: ARG002
         speed: float = 1.0,
     ) -> SynthesisResult:
-        """Synthesize text to audio.
-
-        Args:
-            text: Text to synthesize.
-            voice: Voice to use (not used for Piper - voice is the model).
-            speed: Speech speed multiplier (0.25 to 4.0).
-
-        Returns:
-            SynthesisResult with audio data and metadata.
-
-        Raises:
-            InvalidTextError: If the text is empty or invalid.
-            RuntimeError: If the model is not loaded.
-
-        """
-        if self._voice is None:
-            msg = "Model not loaded"
+        """Synthesize text to audio."""
+        if self._executor is None:
+            msg = "Model not loaded. Call load() first."
             raise RuntimeError(msg)
 
         if not text or not text.strip():
@@ -211,20 +213,14 @@ class PiperBackend:
             raise InvalidTextError(msg)
 
         # Convert speed to length_scale (inverse relationship)
-        # Speed is already validated/clamped by the API layer
         # length_scale < 1.0 = faster, > 1.0 = slower
         length_scale = 1.0 / speed
 
-        # Run synthesis in executor to avoid blocking.
-        # Thread-safe: ONNX Runtime InferenceSession.run() is thread-safe since v1.10+,
-        # so concurrent requests can share the same PiperVoice instance safely.
         loop = asyncio.get_running_loop()
         audio_data, duration = await loop.run_in_executor(
-            None,
-            _synthesize_sync,
-            self._voice,
+            self._executor,
+            _synthesize_in_subprocess,
             text,
-            self._sample_rate,
             length_scale,
         )
 

{agent_cli-0.79.0 → agent_cli-0.80.0}/docs/commands/dev.md

@@ -333,6 +333,7 @@ agent-cli dev clean [OPTIONS]
 | `--no-commits` | `false` | Also remove worktrees with 0 commits ahead of default branch (abandoned branches) |
 | `--dry-run, -n` | `false` | Preview what would be removed without actually removing |
 | `--yes, -y` | `false` | Skip confirmation prompts |
+| `--force, -f` | `false` | Force removal of worktrees with modified or untracked files |
 
 
 <!-- OUTPUT:END -->

{agent_cli-0.79.0 → agent_cli-0.80.0}/docs/installation/macos.md

@@ -52,19 +52,19 @@ The `setup-macos.sh` script:
 - ✅ Installs `uv` if needed
 - ✅ Installs/checks Ollama (native macOS app)
 - ✅ Installs Zellij for session management
-- ✅ Prepares Wyoming service runners
+- ✅ Installs Whisper and TTS as launchd daemons via `agent-cli daemon install`
 
 ## Services Overview
 
-| Service          | Implementation         | Port  | GPU Support          |
-| ---------------- | ---------------------- | ----- | -------------------- |
-| **Ollama**       | Native macOS app       | 11434 | ✅ Metal GPU         |
-| **Whisper**      | Wyoming MLX Whisper    | 10300 | ✅ Apple Silicon MLX |
-| **Piper**        | Wyoming Piper (via uv) | 10200 | N/A                  |
-| **OpenWakeWord** | Wyoming OpenWakeWord   | 10400 | N/A                  |
+| Service          | Implementation          | Port  | GPU Support          |
+| ---------------- | ----------------------- | ----- | -------------------- |
+| **Ollama**       | Native macOS app        | 11434 | ✅ Metal GPU         |
+| **Whisper**      | MLX Whisper (via daemon)| 10300 | ✅ Apple Silicon MLX |
+| **TTS (Kokoro)** | Kokoro TTS (via daemon) | 10200 | ✅ Metal GPU (MPS)   |
+| **OpenWakeWord** | Wyoming OpenWakeWord    | 10400 | N/A                  |
 
 > [!NOTE]
-> Whisper uses [wyoming-mlx-whisper](https://github.com/basnijholt/wyoming-mlx-whisper) with `whisper-large-v3-turbo` for near real-time transcription on Apple Silicon.
+> Whisper and TTS run as launchd daemons via `agent-cli daemon install`, using MLX and Metal for GPU acceleration on Apple Silicon.
 
 ## Session Management with Zellij
 
@@ -95,13 +95,14 @@ brew services start ollama
 # Or run in foreground:
 ollama serve
 
-# Whisper (Apple Silicon: launchd service or manual)
-launchctl list com.wyoming_mlx_whisper
-# Or run in foreground:
-agent-cli server whisper
+# Whisper and TTS are installed as launchd daemons
+agent-cli daemon status              # Check all daemon status
+agent-cli daemon install whisper     # Install whisper daemon
+agent-cli daemon install tts-kokoro  # Install TTS daemon
 
-# Piper
-agent-cli server tts --backend piper
+# Or run in foreground (without daemon):
+agent-cli server whisper
+agent-cli server tts --backend kokoro
 
 # OpenWakeWord
 scripts/run-openwakeword.sh